Differential Effect of Consolidative Thoracic Radiation Therapy in Extensive-Stage Small Cell Lung Cancer Based on Sex

Purpose

The landmark randomized trial on chest irradiation in extensive disease small cell lung cancer (CREST) demonstrated that consolidative thoracic radiation therapy (cTRT) improved overall (OS) and progression-free survival (PFS) after initial chemotherapy (chemo) in extensive-stage small cell lung cancer, with potentially increased benefit in women compared with men. It is unknown whether similar findings would apply after chemoimmunotherapy became the standard first-line treatment. In this analysis, we report national practice patterns and survival outcomes of cTRT according to patient sex.

Methods and Materials

We included patients from de-identified electronic health record-derived database diagnosed with stage IV small cell lung cancer (2014-2021) who completed 4 to 6 cycles of first-line systemic therapy (platinum-doublet chemotherapy or chemoimmunotherapy). We evaluated OS and PFS using multivariable Cox proportional hazards regression with receipt of cTRT as an independent variable and stratified by sex. As a sensitivity analysis, we weighted the models by the inverse probability of receiving cTRT.

Results

A total of 1227 patients were included (850 chemotherapy, 377 chemoimmunotherapy). There were no statistically significant differences in baseline characteristics between patients who did and did not receive cTRT. Among women, cTRT was associated with superior OS (adjusted hazard ratio [HR], 0.67; 95% CI, 0.52-0.87) and PFS (HR, 0.63; 95% CI, 0.49-0.82) compared with those not receiving cTRT. Conversely, no OS or PFS benefit with cTRT was observed in men (OS HR, 1.03; 95% CI, 0.80-1.31; PFS HR, 1.12; 95% CI, 0.85-1.47). Findings were similar in weighted analyses.

Conclusions

The survival efficacy of cTRT may be moderated by sex, with female patients appearing more likely to benefit than male patients. These findings reflect sex-based survival trends with similar effect sizes to those observed in the CREST trial. Although the underpinnings of this association need to be elucidated, stratification by sex should be considered for randomized-controlled trials studying cTRT in extensive-stage small cell lung cancer.

Updated Analysis of Comparative Toxicity of Proton and Photon Radiation for Prostate Cancer

PURPOSE

Previous comparative effectiveness studies have not demonstrated a benefit of proton beam therapy (PBT) compared with intensity-modulated radiation therapy (IMRT) for prostate cancer. An updated comparison of GI and genitourinary (GU) toxicity is needed.

METHODS

We investigated the SEER-Medicare linked database, identifying patients with localized prostate cancer diagnosed from 2010 to 2017. Procedure and diagnosis codes indicative of treatment-related toxicity were identified. As a sensitivity analysis, we also identified toxicity based only on procedure codes. Patients who underwent IMRT and PBT were matched 2:1 on the basis of clinical and sociodemographic characteristics. We then compared GI and GU toxicity at 6, 12, and 24 months after treatment.

RESULTS

The final sample included 772 PBT patients matched to 1,544 IMRT patients. The frequency of GI toxicity for IMRT versus PBT was 3.5% versus 2.5% at 6 months (P = .18), 9.5% versus 10.2% at 12 months (P = .18), and 20.5% versus 23.4% at 24 months (P = .11). The frequency of only procedure codes indicative of GI toxicity for IMRT versus PBT was too low to be reported and not significantly different. The frequency of GU toxicity for IMRT versus PBT was 6.8% versus 5.7% (P = .30), 14.3% versus 12.2% (P = .13), and 28.2% versus 25.8% (P = .21) at 6, 12, and 24 months, respectively. When looking only at procedure codes, the frequency of GU toxicity for IMRT was 1.0% at 6 months, whereas it was too infrequent to report for PBT (P = .64). GU toxicity for IMRT versus PBT was 3.3% versus 2.1% (P = .10), and 8.7% versus 6.7% (P = .10) at 12 and 24 months, respectively.

CONCLUSION

In this observational study, there were no statistically significant differences between PBT and IMRT in terms of GI or GU toxicity.

Incarceration and screen-detectable cancer diagnosis among adults in Connecticut

Although incarcerated adults are at elevated risk of dying from cancer, little is known about cancer screening in carceral settings. This study compared stage-specific incidence of screen-detectable cancers among incarcerated and recently released people with the general population, as a reflection of screening practices. We calculated the age- and sex-standardized incidence ratios (SIR) for early- and late-stage cancers for incarcerated and recently released adults compared to the general Connecticut population between 2005 and 2016. Our sample included 143 cancer cases among those incarcerated, 406 among those recently released, and 201 360 in the general population. The SIR for early-stage screen-detectable cancers was lower among incarcerated (SIR = 0.28, 95% CI = 0.17 to 0.43) and recently released (SIR = 0.69, 95% CI = 0.51 to 0.88) individuals than the general population. Incidence of late-stage screen-detectable cancer was lower during incarceration (SIR = 0.51, 95% CI = 0.27 to 0.88) but not after release (SIR = 1.32, 95% CI = 0.93 to 1.82). Findings suggest that underscreening and underdetection of cancer may occur in carceral settings.

Trends in new and persistent opioid use in older adults with and without cancer

BACKGROUND

The impact of ongoing efforts to decrease opioid use on patients with cancer remains undefined. Our objective was to determine trends in new and additional opioid use in patients with and without cancer.

METHODS

This retrospective cohort study used data from Surveillance, Epidemiology, and End Results program-Medicare for opioid-naive patients with solid tumor malignancies diagnosed from 2012 through 2017 and a random sample of patients without cancer. We identified 238 470 eligible patients with cancer and further focused on 4 clinical strata: patients without cancer, patients with metastatic cancer, patients with nonmetastatic cancer treated with surgery alone ("surgery alone"), and patients with nonmetastatic cancer treated with surgery plus chemotherapy or radiation therapy ("surgery+"). We identified new, early additional, and long-term additional opioid use and calculated the change in predicted probability of these outcomes from 2012 to 2017.

RESULTS

New opioid use was higher in patients with cancer (46.4%) than in those without (6.9%) (P < .001). From 2012 to 2017, the predicted probability of new opioid use was more stable in the cancer strata (relative declines: 0.1% surgery alone; 2.4% surgery+; 8.8% metastatic cancer), than in the noncancer stratum (20.0%) (P < .001 for each cancer to noncancer comparison). Early additional use declined among surgery patients (‒14.9% and ‒17.5% for surgery alone and surgery+, respectively) but was stable among patients with metastatic disease (‒2.8%, P = .50).

CONCLUSIONS

Opioid prescribing declined over time at a slower rate in patients with cancer than in patients without cancer. Our study suggests important but tempered effects of the changing opioid climate on patients with cancer.

Differential Impact of Consolidative Thoracic Radiotherapy in Extensive-Stage Small Cell Lung Cancer Based on Systemic Therapy Type and Sex

PURPOSE/OBJECTIVE(S)

Consolidative thoracic radiotherapy (cTRT) has been shown in phase III RCTs to improve overall (OS) and progression-free survival (PFS) after initial chemotherapy (chemo) in extensive-stage small cell lung cancer (ES-SCLC). This benefit was particularly pronounced in women compared to men in the 2015 CREST trial (hazard ratio [HR] 0.68 vs. 1.01, respectively). However, it is unknown whether similar findings would apply after chemoimmunotherapy (chemo-IO) became standard of care first-line treatment in 2018. In this analysis, we report national practice patterns and survival outcomes of cTRT versus no cTRT following chemo or chemo-IO, stratified by sex.

MATERIALS/METHODS

Patients from the nationwide Flatiron Health de-identified electronic health record-derived database were included if they completed 4-6 cycles of first-line systemic therapy (platinum-doublet chemo or chemo-IO) for stage IV SCLC diagnosed between 2014 and 2021. Patients who progressed or started cTRT within 14 days or died within 90 days of completing systemic therapy were excluded to account for immortal time bias. We evaluated OS and PFS using multivariable Cox proportional hazards regression with receipt of cTRT as an independent covariate and last date of chemo as index date. As a sensitivity analysis to address potential selection bias, we weighted the models by the inverse probability of receiving cTRT. All OS and PFS analyses were stratified by systemic therapy type and sex.

RESULTS

A total of 1,227 patients were included (850 chemo, 377 chemo-IO). The proportion of patients who received cTRT increased from 11.7% in 2014 to 20.7% in 2017, and then decreased to 16.4% in 2021. There were no statistically significant differences in baseline characteristics between patients who did and did not receive cTRT. In adjusted analyses among women receiving chemo, cTRT was associated with superior OS (HR 0.68; 95% confidence interval [CI] 0.51-0.91) and PFS (HR 0.64; 95% CI 0.47-0.86) [Table 1]. There was a non-statistically significant trend towards improved OS (HR 0.57; 95% CI 0.32-1.02) and PFS (HR 0.59; 95% CI 0.34-1.02) among women receiving chemo-IO. No OS or PFS benefit with cTRT was observed in men receiving either chemo or chemo-IO. Findings were similar in weighted analyses.

CONCLUSION

The survival impact of cTRT may be differentially impacted by sex, with female patients appearing more likely to benefit than male patients regardless of systemic therapy type. While the underpinnings of this association need to be elucidated, stratification by sex should be considered for RCTs studying cTRT in ES-SCLC.

Estimating Breast Cancer Overdiagnosis After Screening Mammography Among Older Women in the United States

BACKGROUND

Overdiagnosis is increasingly recognized as a harm of breast cancer screening, particularly for older women.

OBJECTIVE

To estimate overdiagnosis associated with breast cancer screening among older women by age.

DESIGN

Retrospective cohort study comparing the cumulative incidence of breast cancer among older women who continued screening in the next interval with those who did not. Analyses used competing risk models, stratified by age.

SETTING

Fee-for-service Medicare claims, linked to the SEER (Surveillance, Epidemiology, and End Results) program.

PATIENTS

Women 70 years and older who had been recently screened.

MEASUREMENTS

Breast cancer diagnoses and breast cancer death for up to 15 years of follow-up.

RESULTS

This study included 54 635 women. Among women aged 70 to 74 years, the adjusted cumulative incidence of breast cancer was 6.1 cases (95% CI, 5.7 to 6.4) per 100 screened women versus 4.2 cases (CI, 3.5 to 5.0) per 100 unscreened women. An estimated 31% of breast cancer among screened women were potentially overdiagnosed. For women aged 75 to 84 years, cumulative incidence was 4.9 (CI, 4.6 to 5.2) per 100 screened women versus 2.6 (CI, 2.2 to 3.0) per 100 unscreened women, with 47% of cases potentially overdiagnosed. For women aged 85 and older, the cumulative incidence was 2.8 (CI, 2.3 to 3.4) among screened women versus 1.3 (CI, 0.9 to 1.9) among those not, with up to 54% overdiagnosis. We did not see statistically significant reductions in breast cancer-specific death associated with screening.

LIMITATIONS

This study was designed to estimate overdiagnosis, limiting our ability to draw conclusions on all benefits and harms of screening. Unmeasured differences in risk for breast cancer and differential competing mortality between screened and unscreened women may confound results. Results were sensitive to model specifications and definition of a screening mammogram.

CONCLUSION

Continued breast cancer screening was associated with greater incidence of breast cancer, suggesting overdiagnosis may be common among older women who are diagnosed with breast cancer after screening. Whether harms of overdiagnosis are balanced by benefits and for whom remains an important question.

PRIMARY FUNDING SOURCE

National Cancer Institute.

Cancer incidence among incarcerated and formerly incarcerated individuals: A statewide retrospective cohort study

BACKGROUND

Cancer incidence among individuals with incarceration exposure has been rarely studied due to the absence of linked datasets. This study examined cancer incidence during incarceration and postincarceration compared to the general population using a statewide linked cohort.

METHODS

We constructed a retrospective cohort from a linkage of state tumor registry and correctional system data for Connecticut residents from 2005 to 2016, and identified cancers diagnosed during and within 12 months postincarceration. We estimated incidence rates (including for screen-detectable cancers) and calculated the standardized incidence ratios (SIR) for the incarcerated and recently released populations, relative to the general population. We also examined cancer incidence by race and ethnicity within each group.

RESULTS

Cancer incidence was lower in incarcerated individuals (SIR = 0.64, 95% CI 0.56-0.72), but higher in recently released individuals (SIR = 1.34, 95% CI 1.23-1.47) compared with the general population, and across all race and ethnic strata. Similarly, nonscreen-detectable cancer incidence was lower in incarcerated and higher in recently released populations compared to the general population. However, non-Hispanic Black individuals had elevated incidence of screen-detectable cancers compared with non-Hispanic White individuals across all three populations (incarcerated, SIR = 1.66, 95% CI 1.03-2.53; recently released, SIR = 1.83, 95% CI 1.32-2.47; and general population, SIR = 1.18, 95% CI 1.16-1.21).

CONCLUSION

Compared with the general population, incarcerated persons have a lower cancer incidence, whereas recently released persons have a higher cancer incidence. Irrespective of incarceration status, non-Hispanic Black individuals have a higher incidence of screen-detectable cancers compared with non-Hispanic White individuals. Supplemental studies examining cancer screening and diagnoses during incarceration are needed to discern the reasons for observed disparities in incidence.

Clinical and Sociodemographic Factors Associated with Telemedicine Engagement in an Urban Community Health Center Cohort During the COVID-19 Pandemic

Objective: To examine chronic diseases, clinical factors, and sociodemographic characteristics associated with telemedicine utilization among a safety-net population. Materials and Methods: We conducted a retrospective cohort study of adults seeking care in an urban, multisite community health center in the Northeast United States. We included adults with ≥1 outpatient in-person visit during the pre-COVID-19 period (March 1, 2019-February 29, 2020) and ≥1 outpatient visit (in-person or telemedicine) during the COVID-19 period (March 1, 2020-February 29, 2021). Multivariable logistic regression models estimated associations between clinical and sociodemographic factors and telemedicine use, classified as "any" (≥1 visit) and "high" (≥3 visits). Results: Among 5,793 patients who met inclusion criteria, 4,687 (80.9%) had any (≥1) telemedicine visit and 1,053 (18.2%) had high (≥3) telemedicine visits during the COVID-19 period. Older age and Medicare coverage were associated with having any telemedicine use. Older and White patients were more likely to have high telemedicine use. Uninsured patients were less likely to have high telemedicine use. Patients with increased health care utilization in the pre-COVID-19 period and those with hypertension, diabetes, substance use disorders, and depression were more likely to have high telemedicine engagement. Discussion: Chronic conditions, older patients, and White patients compared with Latinx patients, were associated with high telemedicine engagement after adjusting for prior health care utilization. Conclusion: Equity-focused approaches to telemedicine clinical strategies are needed for safety-net populations. Community health centers can adopt disease-specific telemedicine strategies with high patient engagement.

Association Between Age and Survival Trends in Advanced Non-Small Cell Lung Cancer After Adoption of Immunotherapy

Importance

The introduction of immune checkpoint inhibitors (ICIs) has transformed the care of advanced non-small cell lung cancer (NSCLC). Although clinical trials suggest substantial survival benefits, it is unclear how outcomes have changed in clinical practice.

Objective

To assess temporal trends in ICI use and survival among patients with advanced NSCLC across age strata.

Design, Setting, and Participants

This cohort study was performed in approximately 280 predominantly community-based US cancer clinics and included patients aged 18 years or older who had stage IIIB, IIIC, or IV NSCLC diagnosed between January 1, 2011, and December 31, 2019, with follow-up through December 31, 2020. Data were analyzed April 1, 2021, to October 19, 2022.

Main Outcomes and Measures

Median overall survival and 2-year survival probability. The predicted probability of 2-year survival was calculated using a mixed-effects logit model adjusting for demographic and clinical characteristics.

Results

The study sample included 53 719 patients (mean [SD] age, 68.5 [9.3] years; 28 374 men [52.8%]), the majority of whom were White individuals (36 316 [67.6%]). The overall receipt of cancer-directed therapy increased from 69.0% in 2011 to 77.2% in 2019. After the first US Food and Drug Administration approval of an ICI for NSCLC, the use of ICIs increased from 4.7% in 2015 to 45.6% in 2019 (P < .001). Use of ICIs in 2019 was similar between the youngest and oldest patients (aged <55 years, 45.2% vs aged ≥75 years, 43.8%; P = .59). From 2011 to 2018, the predicted probability of 2-year survival increased from 37.7% to 50.3% among patients younger than 55 years and from 30.6% to 36.2% in patients 75 years or older (P < .001). Similarly, median survival in patients younger than 55 years increased from 11.5 months to 16.0 months during the study period, while survival among patients 75 years or older increased from 9.1 months in 2011 to 10.2 months in 2019.

Conclusions and Relevance

This cohort study found that, among patients with advanced NSCLC, the uptake of ICIs after US Food and Drug Administration approval was rapid across all age groups. However, corresponding survival gains were modest, particularly in the oldest patients.

Aggressive End-of-Life Care in the Veterans Health Administration versus Fee-for-Service Medicare among Patients with Advanced Lung Cancer

Background: Unlike fee-for-service Medicare, the Veterans Health Administration (VHA) allows for the provision of concurrent care, incorporating cancer treatment while in hospice. Methods: We compared trends of aggressive care at end of life between Medicare and VHA decedents with advanced nonsmall cell lung cancer from 2006 to 2012, and the relation between regional level end-of-life care between Medicare and VHA beneficiaries. Results: Among 18,371 Veterans and 25,283 Medicare beneficiaries, aggressive care at end of life decreased 15% in VHA and 4% in SEER (Surveillance, Epidemiology, and End Results)-Medicare (p < 0.001). Hospice use significantly increased within both cohorts (VHA 28%-41%; SM 60%-73%, p < 0.001). Veterans receiving care in regions with higher hospice admissions among Medicare beneficiaries were significantly less likely to receive aggressive care at end of life (adjusted odds ratio: 0.13, 95% confidence interval: 0.08-0.23, p < 0.001). Conclusions: Patients receiving lung cancer care in the VHA had a greater decline in aggressive care at end of life, perhaps due to increasing concurrent care availability.

Perioperative magnetic resonance imaging in breast cancer care: Distinct adoption trajectories among physician patient-sharing networks

BACKGROUND

Despite no proven benefit in clinical outcomes, perioperative magnetic resonance imaging (MRI) was rapidly adopted into breast cancer care in the 2000's, offering a prime opportunity for assessing factors influencing overutilization of unproven technology.

OBJECTIVES

To examine variation among physician patient-sharing networks in their trajectory of adopting perioperative MRI for breast cancer surgery and compare the characteristics of patients, providers, and mastectomy use in physician networks that had different adoption trajectories.

METHODS AND FINDINGS

Using the Surveillance, Epidemiology, and End Results-Medicare database in 2004-2009, we identified 147 physician patient-sharing networks (caring for 26,886 patients with stage I-III breast cancer). After adjusting for patient clinical risk factors, we calculated risk-adjusted rate of perioperative MRI use for each physician network in 2004-2005, 2006-2007, and 2008-2009, respectively. Based on the risk-adjusted rate, we identified three distinct trajectories of adopting perioperative MRI among physician networks: 1) low adoption (risk-adjusted rate of perioperative MRI increased from 2.8% in 2004-2005 to 14.8% in 2008-2009), 2) medium adoption (8.8% to 45.1%), and 3) high adoption (33.0% to 71.7%). Physician networks in the higher adoption trajectory tended to have a larger proportion of cancer specialists, more patients with high income, and fewer patients who were Black. After adjusting for patients' clinical risk factors, the proportion of patients undergoing mastectomy decreased from 41.1% in 2004-2005 to 38.5% in 2008-2009 among those in physician networks with low MRI adoption, but increased from 27.0% to 31.4% among those in physician networks with high MRI adoption (p = 0.03 for the interaction term between trajectory group and time).

CONCLUSIONS

Physician patient-sharing networks varied in their trajectory of adopting perioperative MRI. These distinct trajectories were associated with the composition of patients and providers in the networks, and had important implications for patterns of mastectomy use.

Neighborhood and Individual Socioeconomic Disadvantage and Survival Among Patients With Nonmetastatic Common Cancers

IMPORTANCE

Disadvantaged neighborhood-level and individual-level socioeconomic status (SES) have each been associated with suboptimal cancer care and inferior outcomes. However, independent or synergistic associations between neighborhood and individual socioeconomic disadvantage have not been fully examined, and prior studies using simplistic neighborhood SES measures may not comprehensively assess multiple aspects of neighborhood SES.

OBJECTIVE

To investigate the associations of neighborhood SES (using a validated comprehensive composite measure) and individual SES with survival among patients with nonmetastatic common cancers.

DESIGN, SETTING, AND PARTICIPANTS

This prospective, population-based cohort study was derived from the Surveillance, Epidemiology, and End Results-Medicare database from January 1, 2008, through December 31, 2011, with follow-up ending on December 31, 2017. Participants included older patients (≥65 years) with breast, prostate, lung, or colorectal cancer.

EXPOSURES

Neighborhood SES was measured using the area deprivation index (ADI; quintiles), a validated comprehensive composite measure of neighborhood SES. Individual SES was assessed by Medicare-Medicaid dual eligibility (yes vs no), a reliable indicator for patient-level low income.

MAIN OUTCOMES AND MEASURES

The primary outcome was overall mortality, and the secondary outcome was cancer-specific mortality. Hazard ratios (HRs) for the associations of ADI and dual eligibility with overall and cancer-specific mortality were estimated via Cox proportional hazards regression. Statistical analyses were conducted from January 23 to April 15, 2021.

RESULTS

A total of 96 978 patients were analyzed, including 25 968 with breast, 35 150 with prostate, 16 684 with lung, and 19 176 with colorectal cancer. Median age at diagnosis was 76 years (IQR, 71-81 years) for breast cancer, 73 years (IQR, 70-77 years) for prostate cancer, 76 years (IQR, 71-81 years) for lung cancer, and 78 years (IQR, 72-84 years) for colorectal cancer. Among lung and colorectal cancer patients, 8412 (50.4%) and 10 486 (54.7%), respectively, were female. The proportion of non-Hispanic White individuals among breast cancer patients was 83.7% (n = 21 725); prostate cancer, 76.8% (n = 27 001); lung cancer, 83.5% (n = 13 926); and colorectal cancer, 81.1% (n = 15 557). Neighborhood-level and individual-level SES were independently associated with overall mortality, and no interactions were detected. Compared with the most affluent neighborhoods (ADI quintile 1), living in the most disadvantaged neighborhoods (ADI quintile 5) was associated with higher risk of overall mortality (breast: HR, 1.34; 95% CI, 1.26-1.43; prostate: HR, 1.51; 95% CI, 1.42-1.62; lung: HR, 1.21; 95% CI, 1.14-1.28; and colorectal: HR, 1.24; 95% CI, 1.17-1.32). Individual socioeconomic disadvantage (dual eligibility) was associated with higher risk of overall mortality (breast: HR, 1.22; 95% CI, 1.15-1.29; prostate: HR, 1.29; 95% CI, 1.21-1.38; lung: HR, 1.14; 95% CI, 1.09-1.20; and colorectal: HR, 1.23; 95% CI, 1.17-1.29). A similar pattern was observed for cancer-specific mortality.

CONCLUSIONS AND RELEVANCE

In this cohort study, neighborhood-level deprivation was associated with worse survival among patients with nonmetastatic breast, prostate, lung, and colorectal cancer, even after accounting for individual SES. These findings suggest that, in order to improve cancer outcomes and reduce health disparities, policies for ongoing investments in low-resource neighborhoods and low-income households are needed.

Post-operative radiation therapy for non-small cell lung cancer: A comparison of radiation therapy techniques

OBJECTIVES

Post-operative radiation therapy (PORT) in locally advanced non-small cell lung cancer (LA-NSCLC) has historically been associated with toxicity. Conformal techniques like intensity modulated radiation therapy (IMRT) have the potential to reduce acute and long-term toxicity from radiation therapy. Among patients receiving PORT for LA-NSCLC, we identified factors associated with receipt of IMRT and evaluated the association between IMRT and toxicity.

METHODS

We queried the Surveillance, Epidemiology, and End Results (SEER)-Medicare database between January 1, 2006 to December 31, 2014 to identify patients diagnosed with Stage II or III NSCLC and who received upfront surgery and subsequent PORT. Baseline differences between patients receiving 3-dimentional conformal radiation therapy (3D-CRT) and IMRT were assessed using the chi-squared test for proportions and the t-test for means. Multivariable logistic regression was used to identify predictors of receipt of IMRT and pulmonary, esophageal, and cardiac toxicity. Propensity-score matching was employed to reduce the effect of known confounders.

RESULTS

A total of 620 patients met the inclusion criteria, among whom 441 (71.2%) received 3D-CRT and 179 (28.8%) received IMRT. The mean age of the cohort was 73.9 years and 54.7% were male. The proportion of patients receiving IMRT increased from 6.2% in 2006 to 41.4% in 2014 (P < 0.001). IMRT was not associated with decreased pulmonary (OR 0.89; 95% CI, 0.62-1.29), esophageal (OR 1.09; 95% CI, 0.0.75-1.58), or cardiac toxicity (OR 1.02; 95% CI, 0.69-1.51). These findings held on propensity-score matching. Clinical risk factors including comorbidity and prior treatment history were associated with treatment toxicity.

CONCLUSION

In a cohort of elderly patients, the use of IMRT in the setting of PORT for LA-NSCLC was not associated with a difference in toxicity compared to 3D-CRT. This finding suggests that outcomes from PORT may be independent of radiotherapy treatment technique.

Primary anti-infective prophylaxis during routine treatment of lymphoma in the United States: A large real-world cohort analysis

268

Background: While infectious complications contribute to considerable morbidity and mortality in patients with cancer, most scenarios lack evidence to guide optimal anti-infective prophylaxis (AIP). We evaluated a large real-world dataset to identify baseline utilization and factors associated with AIP in patients with non-Hodgkin lymphoma (NHL) treated in the US-community setting. Methods: Using the nationwide Flatiron Health de-identified electronic health record-derived database (from ≈ 280 US cancer clinics), we selected patients treated prior to 7/1/2020 with 1) R-CHOP for DLBCL, 2) bendamustine and rituximab (BR) for CLL/SLL, or 3) ibrutinib for CLL/SLL. We limited our analysis to patients treated by providers with documented prescribing of guideline recommended anti-viral prophylaxis (ppx) during proteasome inhibitor administration to ≥1 multiple myeloma patient. Our main outcome was the documented use of primary AIP defined as anti-viral and/or pneumocystis jiroveci (PJP) ppx within +/- 14 days of treatment initiation. We also report the delayed documented AIP use from day 15 to 60. We applied separate multivariable logistic regression models to each setting to examine the associations of patient-level characteristics with primary AIP (including age, sex, race, region, insurance, ECOG, year of treatment initiation). Results: A total of 3,142 (R-CHOP for DLBCL), 2,180 (BR for CLL/SLL), and 3,590 (ibrutinib for CLL/SLL) patients were included, with median age of 69, 69, and 72 years, respectively. Primary AIP was most common during BR for CLL/SLL, with 16.8% receiving any AIP (antiviral 15.6%, PJP 7.3%). Primary AIP was used in 10.5% of DLBCL patients initiating R-CHOP (antiviral 7.6%, PJP 5.6%), with the lowest utilization of AIP during ibrutinib for CLL/SLL (any 6.4%, antiviral 5.6%, PJP 2.6%). In the delayed setting, an additional 4-6% and 2-5% received viral and PJP ppx, respectively. Across all three of our multivariable analyses, higher provider rate of anti-viral ppx during proteosome inhibitor administration in MM, residing in the Midwest (vs. Northeast), and more recent treatment initiation were associated with greater odds of AIP. Other patient characteristics (age, race, ECOG) were less consistently associated with AIP across models. Furthermore, C-statistics were <0.7 in all three models (0.660-0.685), suggesting suboptimal discrimination for AIP based on patient-level characteristics alone. Conclusions: We observed low utilization of primary AIP during treatment in three common NHL settings that lack clear consensus on AIP. Variation was not well explained by measured patient characteristics, and future studies should consider provider and system attributes. Ultimately, robust evidence generation (e.g. pragmatic clinical trials) and quality improvement measures are needed to optimize ppx during routine lymphoma management.

Temporal Trends in Opioid Prescribing Patterns Among Oncologists in the Medicare Population

BACKGROUND

In the wake of the US opioid epidemic, there have been efforts to curb opioid prescribing. However, it is unknown whether these efforts have affected prescribing among oncologists, whose patients often require opioids for symptom management. We investigated temporal patterns in opioid prescribing for Medicare beneficiaries among oncologists and nononcologists.

METHODS

We queried the Centers for Medicare and Medicaid Services Part D prescriber dataset for all physicians between January 1, 2013, and December 31, 2017. We used population-averaged multivariable negative binomial regression to estimate the association between time and per-provider opioid and gabapentinoid prescribing rate, defined as the annual number of drug claims (original prescriptions and refills) per beneficiary, among oncologists and nononcologists on a national and state level.

RESULTS

From 2013 to 2017, the national opioid-prescribing rate declined by 20.7% (P < .001) among oncologists and 22.8% (P < .001) among non oncologists. During this time frame, prescribing of gabapentin increased by 5.9% (P < .001) and 23.1% (P < .001) among oncologists and nononcologists, respectively. Among palliative care providers, opioid prescribe increased by 15.3% (P < .001). During the 5-year period, 43 states experienced a decrease (P < .05) in opioid prescribing among oncologists, and in 5 states, opioid prescribing decreased more among oncologists than nononcologists (P < .05).

CONCLUSIONS

Between 2013 and 2017, the opioid-prescribing rate statistically significantly decreased nationwide among oncologists and nononcologists, respectively. Given similar declines in opioid prescribing among oncologists and nononcologists, there is concern that opioid-prescribing guidelines intended for the noncancer population are being applied inappropriately to patients with cancer and cancer survivors.

Physician trajectories of abandoning long-course breast radiotherapy and their cost impact

OBJECTIVE

To examine variation in trajectories of abandoning conventionally fractionated whole-breast irradiation (CF-WBI) for adjuvant breast radiotherapy among physician peer groups and the associated cost implications.

DATA SOURCES

Medicare claims data were obtained from the Chronic Conditions Data Warehouse for fee-for-service beneficiaries with breast cancer in 2011-2014.

STUDY DESIGN

We used social network methods to identify peer groups of physicians that shared patients. For each physician peer group in each time period (T1 = 2011-2012 and T2 = 2013-2014), we calculated a risk-adjusted rate of CF-WBI use among eligible women, after adjusting for patient clinical characteristics. We applied a latent class growth analysis to these risk-adjusted rates to identify distinct trajectories of CF-WBI use among physician peer groups. We further estimated potential savings to the Medicare program by accelerating abandonment of CF-WBI in T2 using a simulation model.

DATA COLLECTION/EXTRACTION METHODS

Use of conventionally fractionated whole-breast irradiation was determined from Medicare claims among women ≥ 66 years of age who underwent adjuvant radiotherapy after breast conserving surgery.

PRINCIPAL FINDINGS

Among 215 physician peer groups caring for 16 988 patients, there were four distinct trajectories of abandoning CF-WBI: (a) persistent high use (mean risk-adjusted utilization rate: T1 = 94.3%, T2 = 90.6%); (b) decreased high use (T1 = 81.3%, T2 = 65.3%); (c) decreased medium use (T1 = 60.1%, T2 = 44.0%); and (d) decreased low use (T1 = 31.6%, T2 = 23.6%). Peer groups with a smaller proportion of patients treated at free-standing radiation facilities and a larger proportion of physicians that were surgeons tended to follow trajectories with lower use of CF-WBI. If all physician peer groups had practice patterns in T2 similar to those in the "decreased low use" trajectory, the Medicare program could save $83.3 million (95% confidence interval: $58.5 million-$112.2 million).

CONCLUSIONS

Physician peer groups had distinct trajectories of abandoning CF-WBI. Physician composition and setting of radiotherapy were associated with the different trajectories. Distinct practice patterns across the trajectories had important cost implications.

Adoption of consolidative durvalumab among patients with locally advanced non-small cell lung cancer

e20550

Background: Soon after the publication of the PACIFIC trial, the Food and Drug Administration (FDA) approved durvalumab following definitive chemoradiation in patients with unresectable stage III non-small cell lung cancer (NSCLC). However, patterns of durvalumab adoption within the United States are currently unknown. We investigated the pace of uptake of durvalumab as well as patient and tumor characteristics associated with receipt of durvalumab following FDA approval. Methods: This study used the nationwide Flatiron Health database, a longitudinal electronic health record-derived de-identified database. During the study period, the de-identified data originated from approximately 280 US cancer clinics (̃800 sites of care). From this de-identified database, we selected patients with stage IIIB-C NSCLC who completed at least two cycles of first-line (1L) platinum doublet chemotherapy between February 16, 2018 (FDA approval date for durvalumab) and August 16, 2019, and did not develop disease progression over a follow-up period of at least 120 days after 1L chemotherapy. In the absence of radiation therapy and surgery details in the database, these criteria served as a proxy for selecting patients who were most likely to be eligible for consolidative durvalumab. We evaluated temporal trends in the uptake of durvalumab after completion of 1L chemotherapy using the Cochran-Armitage test, clustering the time periods in 3-month intervals. We also characterized patient and tumor characteristics associated with durvalumab receipt using the chi-square test and multivariable logistic regression. Results: A total of 424 patients were included in our study sample, among whom 282 (66.5%) patients had documentation of having received durvalumab after 1L chemotherapy. Within the first 3 months after FDA approval, 60.4% of potentially eligible patients received durvalumab, while in the final 3 months of the study period, this proportion rose modestly to 68.3% (p = 0.19). On univariable and multivariable analysis, patient characteristics like age, sex, race, insurance, smoking status, performance status, and comorbidity score, as well as tumor characteristics like histology, EGFR status, and PD-L1 status, were not significantly associated with receipt of durvalumab. However, there was significant geographic variability in receipt of durvalumab, ranging from 65.7% in the South to 85.3% in the Midwest (chi-square p = 0.04). Conclusions: Early adoption of durvalumab upon completion of 1L chemotherapy for stage IIIB-C lung cancer was rapid across the United States, starting at over 60% in the first three months after FDA approval. Regional variation appears to be a more powerful driver than patient characteristics in discrepancies of durvalumab dissemination. Further exploration into these regional variations may help improve guideline-concordant care.

Association of Programmed Cell Death Ligand 1 Expression Status With Receipt of Immune Checkpoint Inhibitors in Patients With Advanced Non-Small Cell Lung Cancer

IMPORTANCE

Initial approval for immune checkpoint inhibitors (ICIs) for treatment of advanced non-small cell lung cancer (NSCLC) was limited to patients with high levels of programmed cell death ligand 1 (PD-L1) expression. However, in the period after approval, it is not known how new evidence supporting efficacy of these treatments in patients with low or negative PD-L1 expression was incorporated into real-world practice.

OBJECTIVE

To evaluate the association between PD-L1 testing and first-line ICI use.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective cohort study (January 1, 2011, to December 31, 2018) used a deidentified nationwide electronic health record-derived database reflecting real-world care at more than 280 US community and academic cancer clinics (approximately 800 sites of care). Patients included those with advanced NSCLC without other identifiable variations diagnosed in the period after the US Food and Drug Administration's initial first-line approval of ICIs for patients with high PD-L1 expression (≥50%).

EXPOSURE

First-line ICI treatment.

MAIN OUTCOMES AND MEASURES

Patterns of PD-L1 testing and first-line ICI treatment among all patients and patients stratified by tumor histologic type (squamous vs nonsquamous).

RESULTS

A total of 45 631 patients (mean [SD] age, 68.4 [9.6] years; 21 614 [47.4%] female) with advanced NSCLC were included in the study. PD-L1 testing increased from 468 (7.2%) in 2015 to 4202 (73.2%) in 2018. Within a subset of 7785 patients receiving first-line treatment in the period after first-line approval of pembrolizumab, those who received PD-L1 testing had a greater odds of receiving an ICI (odds ratio, 2.11; 95% CI, 1.89-2.36). Among patients with high PD-L1 expression (≥50%), 1541 (83.5%) received first-line ICI treatment; 776 patients (40.3%) with low PD-L1 expression (1%-49%) and 348 (32.3%) with negative PD-L1 expression (0%) also received ICIs. In addition, 755 untested patients (32.8%) were treated with a first-line ICI. The proportion of patients who received ICIs without PD-L1 testing increased during the study period (59 [17%] in quarter 4 of 2016 to 141 [53.8%] in quarter 4 of 2018).

CONCLUSIONS AND RELEVANCE

In this study, use of first-line ICI treatment increased among patients with advanced NSCLC with negative, low, or untested PD-L1 expression status in 2016 through 2018. These findings suggest that national practice was rapidly responsive to new clinical evidence rather than adhering to regulatory guidance in place at the time.

Temporal trends in opioid prescribing patterns among oncologists in the Medicare population

12022

Background: In the wake of the United States (U.S.) opioid epidemic, there have been significant governmental and societal efforts to curb opioid prescribing. However, it is unknown whether these efforts have affected prescribing among oncologists, whose patient population often requires narcotics for symptom management. We investigated temporal patterns in opioid prescribing for Medicare patients among oncologists. Methods: We queried the Centers for Medicare and Medicaid Services Part D prescriber dataset to identify independently practicing physicians between January 1, 2013 and December 31, 2017. We used population-averaged multivariable negative binomial regression to estimate the association between time and per-provider opioid prescribing rate, defined as number of opioid claims (original prescriptions and refills) per 100 patients, among oncologists and non-oncologists on both a national and statewide level. All models were adjusted for provider characteristics and annual total patient count per provider. Results: The final study sample included 20,513 oncologists and 711,636 non-oncologists. From 2013 to 2017, the national opioid prescribing rate declined by 19.3% (68.8 to 55.5 opioid prescriptions per 100 patients; P< 0.001) among oncologists and 20.4% (50.7 to 40.3 prescriptions per 100 patients; P< 0.001) among non-oncologists. During this timeframe, 40 U.S. states experienced a significant ( P< 0.05) decrease in opioid prescribing among oncologists, most notably in Vermont (-43.2%), Idaho (-34.5%), and Maine (-32.8%). In comparison, all 50 states exhibited a significant decline ( P< 0.05) in opioid prescribing among non-oncologists. In 5 states, opioid prescribing decreased more among oncologists than non-oncologists, including Oklahoma (-24.6% vs. -7.1%), Idaho (-34.5% vs. -17.8%), Utah (-31.7% vs. -18.7%), Texas (-19.9% vs. -14.7%), and New York (-24.0% vs. -19.7%) (all P< 0.05). Conclusions: Between 2013 and 2017, the opioid prescribing rate decreased by approximately 20% nationwide among both oncologists and non-oncologists. These findings raise concerns about whether opioid prescribing legislation and guidelines intended for the non-cancer population are being applied inappropriately to patients with cancer and survivors.

Aggressive care at end-of-life in the Veteran’s Health Administration versus fee-for-service Medicare among patients with advanced lung cancer

12025

Background: The Veteran’s Health Administration (VHA) allows simultaneous receipt of cancer treatment and hospice care, termed concurrent care, while fee-for-service Medicare does not. Although many physicians who care for patients in the VHA also care for private sector patients, it is unclear whether there is a “spillover” relation between end of life (EOL) care in the VHA and Medicare systems at the regional level. We examined temporal trends, as well as regional-level associations between Medicare and VHA EOL practice for patients with advanced lung cancer. Methods: We conducted a retrospective study on VHA and SEER-Medicare (SM) decedents from 2006-2012 with stage IV non-small cell lung cancer (NSCLC) who received any lung cancer care. Aggressive care (AC) at EOL was defined as any of the following within 30 days of death– intensive care unit (ICU) admission, no-hospice care, cardiopulmonary resuscitation(CPR), mechanical ventilation (MV), > 1 inpatient admission and receipt of chemotherapy. Descriptive statistics were used to compare outcomes. We also analyzed the association between Medicare hospital referral region (HRR) hospice admissions, Medicare HRR EOL spending, and VHA AC use adjusted for patient’s characteristics using a random intercept mixed effect logistic regression model after matching VHA facilities with Medicare facilities in a particular HRR. Results: AC use significantly decreased during the study period, from 46% to 31% among 18,371 Veterans and from 42% to 38% among 25,283 in the SM cohort, (t-test P < .05). Hospice use significantly increased within both cohorts (p < .001). The receipt of chemotherapy at EOL was similar for both cohorts throughout the study period. Veterans who received care in regions with higher hospice admissions among Medicare beneficiaries were significantly less likely to receive AC at EOL (adjusted Odds Ratio (aOR): 0.13 95%CI: 0.08-0.23, P < .001) than veterans in regions with lower Medicare hospice use. Medicare HRR spending at the EOL was not associated with receipt of AC among Medicare beneficiaries (aOR): 1.004 95%CI: 1.00-1.009, P = 0.07). Conclusions: Perhaps due to availability of concurrent care, VHA patients received less aggressive care at EOL as compared to SM patients. At the regional level, greater hospice use among Medicare beneficiaries was significantly associated with reduced AC within the VHA.

Hospital Variation in Spending for Lung Cancer Resection in Medicare Beneficiaries

BACKGROUND

As cancer payment models transition from fee for service toward payment "bundles" based on episodes of care, a deeper understanding of the costs associated with stage I lung cancer treatment becomes increasingly relevant. To better understand costs in early lung cancer care, we sought to characterize hospital-level variation in Medicare expenditure after lobectomy for stage I non-small cell lung carcinoma.

METHODS

Patients who were diagnosed with stage I non-small cell lung carcinoma from 2006 through 2011 and undergoing lobectomy were selected from the Surveillance, Epidemiology and End Results-Medicare linked database. We used Medicare claims to estimate costs of care in the 90 days after initial surgical hospitalization. Hospitals were grouped into quintiles of mean excess cost, calculated as the mean difference between observed costs and risk-adjusted predicted costs. The association between hospital factors and mean excess cost were compared across hospitals, including complication rates and hospital volume.

RESULTS

A total of 3530 patients underwent lobectomy at 156 hospitals. Hospitals in the lowest cost quintile had index hospitalizations $6226 less costly than predicted. Conversely, the most expensive hospital quintile had index hospital costs that were $6151 costlier than predicted. Increased costs were positively associated with the number of complications per patient (P < .001), but not hospital volume (P = .85).

CONCLUSIONS

Among Medicare beneficiaries undergoing lobectomy for stage I non-small cell lung carcinoma, the cost of perioperative care varied substantially across hospitals and was strongly associated with complication rate, but not hospital volume.

Persistent Use of Extended Fractionation Palliative Radiotherapy for Medicare Beneficiaries With Metastatic Breast Cancer, 2011 to 2014

INTRODUCTION

With no evidence to support extended radiation courses for the palliation of bone metastases, multiple guidelines were issued discouraging its use. We assessed contemporary use and cost of prolonged palliative radiotherapy in Medicare beneficiaries with bone metastases from breast cancer.

METHODS

We conducted a retrospective, longitudinal study of palliative radiotherapy use among fee-for-service Medicare beneficiaries with bone metastasis from breast cancer who underwent palliative radiotherapy during 2011 to 2014. Patients were categorized according to the number of days (fractions) on which they received palliative radiotherapy: 1, 2 to 10, 11 to 19, or 20 to 30. We examined the association of clinical, demographic, and provider characteristics with the use of extended (≥11 fractions) or very extended (≥20 fractions) fractionation with logistic regression models. We also compared the cost of different fractionation schemes from the payer perspective.

RESULTS

Of the 7547 patients in the sample (mean age, 71 y), 3084 (40.8%) received extended fractionation. The proportion of patients receiving 11 to 19 (34.7% in 2011 and 28.1% in 2014, trend P<0.001) and 20 to 30 treatments (10.3% in 2011 to 9.0% in 2014, trend P=0.07) decreased modestly over time. Patients with comorbidities were less likely to undergo extended fractionation (34.4% for ≥3 comorbidities vs. 44.9% for 0 comorbidities; adjusted odds ratio 0.67 [95% confidence interval, 0.58-0.76]). Patients treated at free-standing practices were more likely to undergo extended fractionation (47.9%) compared with those treated at hospital-based practices (37.3%, P<0.001; adjusted odds ratio, 1.49 [95% confidence interval, 1.35-1.65]). The mean cost of treatment varied from $633 (SD $240) for single-fraction treatment, to $3566 (SD $1349) for 11 to 19 fractions, to $6597 (SD $2893) for 20 to 30 fractions.

CONCLUSION

The use of prolonged courses of palliative radiotherapy among Medicare beneficiaries with breast cancer remained high in 2011 to 2014. The association between free-standing facility status and use of extended fractionation suggests that provider financial incentives may impact choice of treatment.

Real-world practice patterns and impact of PD-L1 expression testing in patients with advanced non-small cell lung cancer

9059

Background: Several tests measuring programmed cell death ligand 1 (PD-L1) expression are available to select patients with advanced non-small cell lung cancer (aNSCLC) for PD-1 axis therapy including FDA approved companion diagnostics (CoDx), as well as complementary (CyDx) and laboratory developed tests (LDTs). However, it is unknown how rapidly PD-L1 expression testing has become adopted, which type of tests are used, or whether testing affects first-line therapy. Methods: We retrospectively reviewed the Flatiron Health electronic health record-derived database, reflecting real-world care at community oncology practices in the U.S. We evaluated trends in the use of PD-L1 expression testing within 12 months of diagnosis. Among patients diagnosed after FDA approval of PD-L1 testing prior to first-line therapy (Oct 2016), and without other identifiable driver mutations, we used multivariable logistic regression to examine the association between PD-L1 testing and receipt of first-line PD-1 axis therapy. We further examined the association between PD-L1 expression result (using overall reported status or % staining) and choice of first-line therapy. Results: We identified 49,546 patients with aNSCLC from 2011 through the third quarter of 2018. The use of PD-L1 expression testing increased from 7.0% in 2015 to 70.6% in 2017 (p < 0.01). Among those receiving testing, the proportion of patients receiving CoDx increased from 29.9% in 2015 to 73.0% in 2017 with corresponding reductions in CyDx and LDTs. PD-L1 expression testing was associated with use of PD1-axis therapy in the first-line (OR = 2.92, 95% CI 2.59-3.30). Among 4,942 treated patients diagnosed after Oct 2016, those with ≥50% staining more frequently received first line PD-1 axis therapy (82.9%) than those with intermediate (35.5%) or low (25.8%) positive staining; conversely, patients with ≥50% were less likely to receive chemotherapy alone or combination with PD-1 axis therapy. Conclusions: PD-L1 expression testing was rapidly adopted following FDA approval of companion diagnostic testing for aNSCLC. Although the results of PD-L1 expression testing inform the choice of first-line therapy, a substantial proportion of patients are not tested prior to first line treatment.

Abandonment trajectories of conventionally fractionated adjuvant radiotherapy in breast cancer care

531

Background: Hypofractionated radiotherapy is the recommended approach for adjuvant breast cancer care. Yet physicians have been slow to abandon conventionally fractionated radiotherapy (CFRT). We examined distinct trajectories of abandoning CFRT among physician patient-sharing peer groups and the associated cost implications. Methods: Using 2011-2014 national Medicare claims, we constructed peer groups of physicians (radiation and medical oncologists, surgeons, and primary care physicians) who cared for women with breast cancer. Women ≥66 years of age who underwent lumpectomy plus adjuvant radiotherapy were included. Peer groups represented physicians who frequently shared patients with one another. For each peer group, we calculated risk-adjusted rate of CFRT use in 2011-2012 (T1) and 2013-2014 (T2) after accounting for patient risk factors. Based on these utilization rates and a latent growth curve analysis, we identified distinct trajectories of abandoning CFRT among peer groups and estimated their cost implications from the Medicare perspective. Results: The 215 physician peer groups (caring for 16,988 patients) exhibited four distinct trajectories of CFRT use: 1) persistent high use (mean adjusted utilization: T1 = 94.3%, T2 = 90.6%); 2) decreased high use (T1 = 81.3%, T2 = 65.3%); 3) decreased medium use (T1 = 60.1%, T2 = 44.0%); and 4) decreased low use (T1 = 31.6%, T2 = 23.6%). They accounted for 33.0%, 35.3%, 25.6% and 6.0% of the peer groups, respectively. Compared to “persistent high use” of CFRT, peer groups with “decreased high use” and “decreased medium use” of CFRT had a smaller proportion of patients receiving radiotherapy at free-standing (vs. hospital-based) facilities (adjusted odds ratio = 0.89, p = 0.01; and 0.78, p < 0.01; respectively). Accelerating abandonment of CFRT in the three higher utilization trajectories to “decreased low use” for the 2011-2014 patient cohort could save Medicare $342.7 million (95% confidence interval: $232.2-$457.6 million). Conclusions: Physician peer groups had distinct trajectories in abandoning CFRT exhibiting different baseline levels and rates of change in utilization. Efforts to reduce overuse of CFRT could generate substantial savings.

Association of State Dense Breast Notification Laws With Supplemental Testing and Cancer Detection After Screening Mammography

OBJECTIVES

To evaluate the association of state dense breast notification (DBN) laws with use of supplemental tests and cancer diagnosis after screening mammography.

METHODS

We examined screening mammograms (n = 1 441 544) performed in 2014 and 2015 among privately insured women aged 40 to 59 years living in 9 US states that enacted DBN laws in 2014 to 2015 and 25 US states with no DBN law in effect. DBN status at screening mammography was categorized as no DBN, generic DBN, and DBN that mandates notification of possible benefits of supplemental screening (DBN+SS). We used logistic regression to examine the change in rate of supplemental ultrasound, magnetic resonance imaging, breast biopsy, and breast cancer detection.

RESULTS

DBN+SS laws were associated with 10.5 more ultrasounds per 1000 mammograms (95% CI = 3.0, 17.6 per 1000; P = .006) and 0.37 more breast cancers detected per 1000 mammograms (95% CI = 0.05, 0.69 per 1000; P = .02) compared with no DBN law. No significant differences were found for generic DBN laws in either ultrasound or cancer detection.

CONCLUSIONS

DBN legislation is associated with increased use of ultrasound and cancer detection after implementation only when notification of the possible benefits of supplemental screening is required.

An empiric approach to identifying physician peer groups from claims data: An example from breast cancer care

OBJECTIVE

To develop an empiric approach for evaluating the performance of physician peer groups based on patient-sharing in administrative claims data.

DATA SOURCES

Surveillance, Epidemiology and End Results-Medicare linked dataset.

STUDY DESIGN

Applying social network theory, we constructed physician peer groups for patients with breast cancer. Under different assumptions of key parameter values-minimum patient volume for physician inclusion and minimum number of patients shared between physicians for a connection-we compared agreement in group membership between split samples during 2004-2006 (T1) (reliability) and agreement in group membership between T1 and 2007-2009 (T2) (stability). We also compared the results with those derived from randomly generated groups and to hospital affiliation-based groups.

PRINCIPAL FINDINGS

The sample included 142 098 patients treated by 43 174 physicians in T1 and 136 680 patients treated by 51 515 physicians in T2. We identified parameter values that resulted in a median peer group reliability of 85.2 percent (Interquartile range (IQR) [0 percent, 96.2 percent]) and median stability of 73.7 percent (IQR [0 percent, 91.0 percent]). In contrast, stability of randomly assigned peer groups was 6.2 percent (IQR [0 percent, 21.0 percent]). Median overlap of empirical groups with hospital groups was 32.2 percent (IQR [12.1 percent, 59.2 percent]).

CONCLUSIONS

It is feasible to construct physician peer groups that are reliable, stable, and distinct from both randomly generated and hospital-based groups.

Surgeon peer network characteristics and adoption of new imaging techniques in breast cancer: A study of perioperative MRI

Background

Perioperative MRI has disseminated into breast cancer practice despite equivocal evidence. We used a novel social network approach to assess the relationship between the characteristics of surgeons’ patient‐sharing networks and subsequent use of MRI.

Methods

We identified a cohort of female patients with stage 0‐III breast cancer from the Surveillance, Epidemiology, and End Results (SEER)‐Medicare database. We used claims data from these patients and non‐cancer patients from the 5% Medicare sample to identify peer groups of physicians who shared patients during 2004‐2006 (T1). We used a multivariable hierarchical model to identify peer group characteristics associated with uptake of MRI in T2 (2007‐2009) by surgeons who had not used MRI in T1.

Results

Our T1 sample included 15 149 patients with breast cancer, treated by 2439 surgeons in 390 physician groups. During T1, 9.1% of patients received an MRI; the use of MRI varied from 0% to 100% (IQR 0%, 8.5%) across peer groups. After adjusting for clinical characteristics, patients treated by surgeons in groups with a higher proportion of primary care physicians (PCPs) in T1 were less likely to receive MRI in T2 (OR = 0.81 for 10% increase in PCPs, 95% CI = 0.71, 0.93). Surgeon transitivity (ie, clustering of surgeons) was significantly associated with MRI receipt (P = 0.013); patients whose surgeons were in groups with higher transitivity in T1 were more likely to receive MRI in T2 (OR = 1.29 for 10% increase in clustering, 95% CI = 1.06, 1.58).

Conclusion

The characteristics of a surgeon's peer network are associated with their patients’ subsequent receipt of perioperative MRI.

Association of Broad-Based Genomic Sequencing With Survival Among Patients With Advanced Non-Small Cell Lung Cancer in the Community Oncology Setting

IMPORTANCE

Broad-based genomic sequencing is being used more frequently for patients with advanced non-small cell lung cancer (NSCLC). However, little is known about the association between broad-based genomic sequencing and treatment selection or survival among patients with advanced NSCLC in a community oncology setting.

OBJECTIVE

To compare clinical outcomes between patients with advanced NSCLC who received broad-based genomic sequencing vs a control group of patients who received routine testing for EGFR mutations and/or ALK rearrangements alone.

DESIGN, SETTING, AND PARTICIPANTS

Retrospective cohort study of patients with chart-confirmed advanced NSCLC between January 1, 2011, and July 31, 2016, and who received care at 1 of 191 oncology practices across the United States using the Flatiron Health Database. Patients were diagnosed with stage IIIB/IV or unresectable nonsquamous NSCLC who received at least 1 line of antineoplastic treatment.

EXPOSURES

Receipt of either broad-based genomic sequencing or routine testing (EGFR and/or ALK only). Broad-based genomic sequencing included any multigene panel sequencing assay examining more than 30 genes prior to third-line treatment.

MAIN OUTCOMES AND MEASURES

Primary outcomes were 12-month mortality and overall survival from the start of first-line treatment. Secondary outcomes included frequency of genetic alterations and treatments received.

RESULTS

Among 5688 individuals with advanced NSCLC (median age, 67 years [interquartile range, 41-85], 63.6% white, 80% with a history of smoking); 875 (15.4%) received broad-based genomic sequencing and 4813 (84.6%) received routine testing. Among patients who received broad-based genomic sequencing, 4.5% received targeted treatment based on testing results, 9.8% received routine EGFR/ALK targeted treatment, and 85.1% received no targeted treatment. Unadjusted mortality rates at 12 months were 49.2% for patients undergoing broad-based genomic sequencing and 35.9% for patients undergoing routine testing. Using an instrumental variable analysis, there was no significant association between broad-based genomic sequencing and 12-month mortality (predicted probability of death at 12 months, 41.1% for broad-based genomic sequencing vs 44.4% for routine testing; difference -3.6% [95% CI, -18.4% to 11.1%]; P = .63). The results were consistent in the propensity score-matched survival analysis (42.0% vs 45.1%; hazard ratio, 0.92 [95% CI, 0.73 to 1.11]; P = .40) vs unmatched cohort (hazard ratio, 0.69 [95% CI, 0.62 to 0.77]; log-rank P < .001).

CONCLUSIONS AND RELEVANCE

Among patients with advanced non-small cell lung cancer receiving care in the community oncology setting, broad-based genomic sequencing directly informed treatment in a minority of patients and was not independently associated with better survival.

Physician peer group characteristics and timeliness of breast cancer surgery

PURPOSE

Little is known about how the structure of interdisciplinary groups of physicians affects the timeliness of breast cancer surgery their patients receive. We used social network methods to examine variation in surgical delay across physician peer groups and the association of this delay with group characteristics.

METHODS

We used linked Surveillance, Epidemiology, and End Results-Medicare data to construct physician peer groups based on shared breast cancer patients. We used hierarchical generalized linear models to examine the association of three group characteristics, patient racial composition, provider density (the ratio of potential vs. actual connections between physicians), and provider transitivity (clustering of providers within groups), with delayed surgery.

RESULTS

The study sample included 8338 women with breast cancer in 157 physician peer groups. Surgical delay varied widely across physician peer groups (interquartile range 28.2-50.0%). For every 10% increase in the percentage of black patients in a peer group, there was a 41% increase in the odds of delayed surgery for women in that peer group regardless of a patient's own race [odds ratio (OR) 1.41, 95% confidence interval (CI) 1.15-1.73]. Women in physician peer groups with the highest provider density were less likely to receive delayed surgery than those in physician peer groups with the lowest provider density (OR 0.65, 95% CI 0.44-0.98). We did not find an association between provider transitivity and delayed surgery.

CONCLUSIONS

The likelihood of surgical delay varied substantially across physician peer groups and was associated with provider density and patient racial composition.

Quality Of Breast Cancer Care In The US Territories: Insights From Medicare

The quality of breast cancer care among Medicare beneficiaries in the US territories-where federal spending for health care is lower than in the continental US-is unknown. We compared female Medicare beneficiaries who were residents of the US territories and had surgical treatment for breast cancer in 2008-14 to those in the continental US in terms of receipt of recommended breast cancer care (diagnostic needle biopsy and adjuvant radiation therapy [RT] following breast-conserving surgery) and the timeliness (time from needle biopsy to surgery and from surgery to adjuvant RT) of that care. Residents of the US territories were less likely to receive recommended care (24 percent lower odds of receiving diagnostic needle biopsy and 34 percent lower odds of receiving adjuvant RT) and to receive timely care (45 percent lower odds of receiving surgery and 82 percent lower odds of receiving adjuvant RT, both within three months). Further research is needed to identify barriers to the provision of adequate and timely breast cancer care in this unique population.

Impact of Widespread Cervical Cancer Screening: Number of Cancers Prevented and Changes in Race-specific Incidence

OBJECTIVES

With recent approval of standalone HPV testing and increasing uptake of HPV vaccination, some have postulated that we are moving toward a "post-Pap" era of cervical cancer prevention. However, the total number cases that have been prevented by Pap smear screening as well as its impact on racial disparities are unknown.

METHODS

We estimated national cervical cancer incidence from 1976 to 2009 using the Surveillance, Epidemiology, and End Result database. Screening data were obtained from the literature and National Cancer Institute Progress Reports. We examined early, late, and race-specific trends in cancer incidence, and calculated the estimated number of cancers prevented over the past 3 decades.

RESULTS

From 1976 to 2009, there was a significant decrease in the incidence of early-stage cervical cancer, from 9.8 to 4.9 cases per 100,000 women (P<0.001). Late-stage disease incidence also decreased, from 5.3 to 3.7 cases per 100,000 women (P<0.001). The incidence among black women decreased from 26.9 to 9.7 cases per 100,000 women (P<0.001), a greater decline compared with that of white women and women of other races. After adjusting for "prescreening era" rates of cervical cancer, we estimate that Pap smears were associated with a reduction of between 105,000 and 492,000 cases of cervical cancer over the past 3 decades in the United States.

CONCLUSIONS

A large number of early-stage and late-stage cervical cancers were prevented and racial disparity in cancer rates were reduced during an era of widespread Pap smear screening.

The Impact of Social Contagion on Physician Adoption of Advanced Imaging Tests in Breast Cancer

Background

Magnetic resonance imaging (MRI) and positron emission tomography (PET) scans are widely used in breast cancer practice despite unproven benefits. We examined the extent to which social contagion is associated with adoption of these imaging modalities.

Methods

We used Surveillance, Epidemiology, and End Results-Medicare to construct peer groups of physicians who shared patients during a baseline period when these imaging modalities were starting to disseminate into practice (2004-2006) and determined the potential impact of these peer groups during a follow-up period (2007-2009). For non-early-adopting surgeons (whose patients did not receive MRI/PET during baseline), we used hierarchical logistic regression models to examine the effect of their peer group's baseline use on their use of MRI/PET during the follow-up period, adjusting for patient characteristics and hospital MRI/PET use.

Results

For MRI, there were 6424 women diagnosed in the follow-up period assigned to 986 non-early-adopting surgeons. During baseline, 9.3% of women received an MRI, varying across peer groups from 0% to 81%. Women assigned to surgeons whose peers had the highest rate of baseline MRI use were more likely to receive MRI compared with women whose surgeons' peers did not use MRI (24.9% vs 10.1%, adjusted odds ratio [OR] = 2.47, 95% confidence interval [CI] = 1.39 to 4.39). Physician peers were associated with uptake of PET imaging (OR for highest vs lowest baseline peer group PET use = 2.04, 95% CI = 1.24 to 3.36).

Conclusions

The phenomenon of social contagion may offer opportunities to better understand how new approaches to cancer care disseminate into clinical practice.

Radiation dose and cardiac risk in breast cancer treatment: An analysis of modern radiation therapy including community settings

PURPOSE

Adjuvant radiation therapy (RT) for breast cancer improves outcomes, but prior studies have documented substantive cardiac dose and cardiac risk. We assessed the mean heart dose (MHD) of RT and estimated the risk of RT-associated cardiac toxicity in women undergoing adjuvant RT for breast cancer in contemporary (predominantly) community practice.

METHODS AND MATERIALS

We identified women with left-sided breast cancer receiving adjuvant RT between 2012 and 2014 from 94 centers across 16 states. We used bivariate analyses and multivariable linear regression to assess associations between RT techniques and MHD. Excess RT-related cardiac risk by age 80 was estimated for women diagnosed at age 60 using the previously reported relationship between MHD and cardiac risk.

RESULTS

Among 1161 women, 77.3% were treated in community practice and with breast conservation (77.8%). The most common techniques were free-breathing (92.2%), supine (94.8%), and fixed gantry intensity modulated RT (FG-IMRT; 46.9%). The median MHD was 2.76 Gy (interquartile range, 1.47-5.03). In multivariable analyses, the predicted median MHD with deep inspiration breath hold was 2.41 Gy compared with 3.86 Gy with free-breathing (P < .001). Three-dimensional conformal RT (3D-CRT) was associated with a lower predicted median MHD (2.78 Gy) than FG-IMRT (4.02 Gy) or rotational IMRT, 6.60 Gy, P < .001). For 60-year-old women with the median MHD of the study population (2.76 Gy) and no cardiovascular risk factors, the 20-year predicted excess risk of death from ischemic heart disease attributable to radiation was 3.5 excess events/1000 patients, in contrast to estimates of 8 events/1000 from prior analyses. The predicted risk of cardiac events varied based on radiation technique, with 4 excess events/1000 with 3D-CRT, 5 excess events/1000 with FG-IMRT, and 8 excess events/1000 with rotational IMRT.

CONCLUSIONS

MHD varies substantially across patients and is influenced by technique in predominantly community settings. Overall risk of cardiac toxicity is modest.

Increased Number of Beam Angles Is Associated With Higher Cardiac Dose in Adjuvant Fixed Gantry Intensity Modulated Radiation Therapy of Left-Sided Breast Cancer

PURPOSE

To analyze the relationship between angle number and mean heart dose (MHD) in adjuvant fixed gantry intensity modulated radiation therapy (FG-IMRT) treatment of left-sided breast cancer as is currently practiced in the community.

METHODS AND MATERIALS

We performed a retrospective, multi-institutional review of women with left-sided breast cancer receiving adjuvant FG-IMRT between 2012 and 2014, encompassing 85 centers in 15 states. Bivariate and multivariate regression analyses were done to identify factors associated with MHD. Long-term cardiac risk was estimated according to a previously published model.

RESULTS

Of the 538 women included, 284 had >2 gantry angle treatment plans (multi-angle), and 254 had 2 gantry angle (standard) plans. Median MHD was higher in patients with multi-angle plans compared with standard (median 475 vs 203 cGy). Number of gantry angles was significantly associated with MHD, with multi-angle plans independently increasing MHD by 229 cGy. Absolute risk of acute coronary events 20 years after treatment was estimated as 7 excess events per 1000 women for standard plans, compared with 12 excess events for multi-angle plans.

CONCLUSIONS

Fixed gantry IMRT breast treatment plans with >2 gantry angles were associated with increased MHD, which translated to an increased cardiac risk. Clinicians should account for this potential drawback in treatment technique when assessing overall plan quality.

Duration of neoadjuvant endocrine therapy and breast cancer outcomes

e12129

Background: Interest in the use of neoadjuvant endocrine therapy (NET) has increased over the past decade, and recent studies have demonstrated similar response rates compared to neoadjuvant chemotherapy. We examined the associations between duration of NET and type of surgery and change from clinical stage to pathologic stage. Methods: We used the National Cancer Data Base to identify women diagnosed with stage II-III, ER and/or PR positive breast cancer who received endocrine therapy from 2004-2014 and underwent surgery. We classified patients according to timing and duration of NET. We performed logistic regression to examine the impact of NET duration on likelihood of 1) receiving breast conserving surgery (BCS) versus mastectomy and 2) being downstaged. Downstaging was defined as pathologic stage lower than clinical stage and upstaging as pathologic stage higher than clinical stage. Results: In our sample of 159,676 patients, 6584 received NET (4.1%). NET was more frequently used in older women with multiple comorbid conditions, larger tumors, and higher nodal stage. The highest rates of NET were in academic/research or integrated network cancer programs. Of patients who underwent NET, 26.5% received it for ≤3 months, 41.0% for 3-6 months, 27.7% for 6-12 months, and 4.9% for 12-24 months. Patients who received NET for 3-6, 6-12, and 12-24 months were all significantly more likely to receive BCS versus mastectomy than patients who did not receive NET (Table). Patients who underwent NET compared to adjuvant endocrine therapy were more likely to be downstaged (Table). Conclusions: The most common duration of NET was 3-6 months, but >26% of patients received NET for less than 3 months. Longer durations of NET were strongly associated with BCS and downstaging. Further research is needed to determine the population of patients mostly likely to benefit from NET and the optimal length of treatment. [Table: see text]

Disparities in next generation sequencing in a population-based community cohort of patients with advanced non-small cell lung cancer

6563

Background: The use of next generation sequencing (NGS) in patients with advanced non-small cell lung cancer (NSCLC) is increasing. This study explored disparities in the use of NGS testing. Methods: This retrospective observational study utilized Flatiron Health’s longitudinal, demographically and geographically diverse database containing electronic health record data from 191 oncology practices across the U.S. We identified patients diagnosed with advanced (stages IIIB/IV or recurrent) non-squamous NSCLC who received first line treatment and either NGS testing or standard biomarker testing (e.g., EGFR, ALK) alone. NGS included any multi-gene panel testing > 30 genes. Logistic regression modeled the association between patient characteristics and receipt of NGS testing, accounting for clustering of patients by oncology practice. Results: Among 5,688 adults with advanced NSCLC, 4,813 (84.6%) patients received standard biomarker testing alone and 875 (15.4%) patients received NGS testing. The median age of the sample was 67y (IQR: 41-85), the majority was white (63.6%) vs. black (7.5%) vs. unknown (13.4%), and had a history of smoking (79.9%). Among the youngest patients ( < 45y), 31.5% received NGS compared to 11.3% among the oldest (76-85y; P < .001). Approximately 16% of white patients received testing, compared to 11.4% of black patients (P < .001). Patients with Medicaid received testing less often than commercially insured patients (11.7% vs 17.0%; P = .10). Patients had significantly lower odds of receiving NGS testing if they were older (≥75 vs. < 45 years of age; adjusted OR: 0.21, 95% CI: 0.13-0.34), black vs. white race (aOR: 0.63, 95% CI: 0.44-0.90) or were Medicaid vs. commercially insured (aOR: 0.54, 95% CI: 0.30-0.97). Conclusions: Significant age, race, and insurance-related disparities exist in the receipt of NGS testing among patients with advanced lung cancer in real world clinical practice.

Myelodysplastic Syndromes and Acute Myeloid Leukemia After Radiotherapy for Prostate Cancer: A Population-Based Study

BACKGROUND

To understand the impact of radiotherapy on the development of myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) among elderly prostate cancer patients.

METHODS

We performed a retrospective cohort study of elderly prostate cancer patients diagnosed during 1999-2011 by using the National Cancer Institute's Surveillance, Epidemiology and End Results-Medicare linked database. Competing risk analyses adjusting for patient characteristics were conducted to assess the impact of radiotherapy on the development of subsequent MDS/AML, compared with surgery.

RESULTS

Of 32,112 prostate cancer patients, 14,672 underwent radiotherapy, and 17,440 received surgery only. The median follow-up was 4.68 years. A total of 157 (0.47%) prostate cancer patients developed subsequent MDS or AML, and the median time to develop MDS/AML was 3.30 (range: 0.16-9.48) years. Compared with prostate cancer patients who received surgery only, patients who underwent radiotherapy had a significantly increased risk of developing MDS/AML (hazard ratio [HR] =1.51, 95% confidence interval [CI]: 1.07-2.13). When radiotherapy was further categorized by modalities (brachytherapy, conventional conformal radiotherapy, and intensity-modulated radiotherapy [IMRT]), increased risk of second MDS/AML was only observed in the IMRT group (HR = 1.66, 95% CI: 1.09-2.54).

CONCLUSIONS

Our findings suggest that radiotherapy for prostate cancer increases the risk of MDS/AML, and the impact may differ by modality. Additional studies with longer follow-up are needed to further clarify the role of radiotherapy in the development of subsequent myeloid malignancies. A better understanding may help patients, physicians, and other stakeholders make more informed treatment decisions. Prostate 77:437-445, 2017. © 2016 Wiley Periodicals, Inc.

Bone Density Screening in Postmenopausal Women With Early-Stage Breast Cancer Treated With Aromatase Inhibitors

PURPOSE

In postmenopausal women with breast cancer treated with aromatase inhibitors (AIs), most expert panels advise baseline bone mineral density testing with a dual-energy x-ray absorptiometry (DXA) scan repeated every 1 to 2 years. How often this recommendation is followed is unclear.

METHODS

We performed a retrospective analysis of women with stage I to III breast cancer who started AI therapy from January 1, 2008, to December 31, 2010, with follow-up through December 31, 2012, by using the SEER-Medicare database. Selection criteria included AI use for ≥ 6 months and no recent osteoporosis diagnosis or bisphosphonate use. We used multivariable logistic regression to investigate associations between patient characteristics and receipt of a baseline DXA scan. In patients who continued AI treatment, we assessed rates of follow-up scans.

RESULTS

In the sample of 2,409 patients (median age, 74 years), 51.0% received a baseline DXA scan. Demographic characteristics associated with the absence of a baseline DXA scan were older age (85 to 94 years v 67 to 69 years; odds ratio [OR], 0.62; 95% CI, 0.42 to 0.92) and black v white race (OR, 0.68; 95% CI, 0.47 to 0.97). Among patients who underwent a baseline DXA scan and continued AI for 3 years, 28.0% had a repeat DXA scan within 2 years and 65.9% within 3 years. In aggregate, of the 1,164 patients who continued with AI treatment for 3 years, only 34.5% had both a baseline and at least one DXA scan during the 3-year follow-up period.

CONCLUSION

The majority of older Medicare beneficiaries with breast cancer treated with AIs do not undergo appropriate bone mineral density evaluation.

Abstract P3-10-03: Socioeconomic disparities in needle biopsy prior to breast cancer surgery across physician referral networks

Introduction

Although needle biopsy (NB) is recommended prior to breast cancer surgery, the use of NB has been shown to vary according to patient socioeconomic status (SES), operating surgeon, and geographic region. We hypothesized that surgeons who work in the same peer referral network (defined by patient sharing) might have similar practice patterns with regard to NB, and that the magnitude of SES disparities might vary across networks. We therefore examined: 1) SES disparities in the receipt of NB, 2) variation in NB across networks, and 3) whether the association between SES and NB varied across networks.

Methods

We used the SEER database and 5% Medicare sample to examine all patients with a new diagnosis of breast cancer from 2004 through 2006. We used Medicare claims to construct peer groups of physicians based on patient-sharing ties. Patients were assigned to peer groups based on the surgeon who performed their definitive surgery. We defined a patient as having low SES if she was in the lowest quintile of area-level income. We used hierarchical generalized linear models (HGLM) to assess the association between low SES and receipt of NB, including random effects for the surgeon, peer group, and Hospital Referral Region (HRR). We then allowed the low SES effect to vary across peer groups in order to determine whether the association between SES and NB varied across groups.

Results

In the full sample of 14,552 patients, 9,498 (65%) received needle biopsy. In bivariable analysis, patients in the lowest income quintile were less likely to receive NB compared to all other patients (59% vs 67%, p&lt;.001). The majority of the variance (59%) in NB use was at the patient level, 22% was at the surgeon level, and 13.7% at the peer group level. The use of NB varied substantially across peer groups, with a median of 69% (interquartile range [51%, 84%]). Even after accounting for physician, peer group, and HRR variation, patients in the lowest stratum of SES were significantly less likely to have received NB compared to all other patient (OR = 0.88; p=.04). Finally, we found that the association between SES and NB varied significantly across referral networks (P&lt;0.05)

Conclusions

Patients with low SES are significantly less likely to receive NB prior to breast cancer surgery, and moreover the magnitude of this SES-related disparity varies significantly according to which referral networks are providing care. Future policies to increase NB rates and standardize care for all breast cancer patients may consider the implications of how care for patients with low SES varies across surgical provider networks.Introduction

Although needle biopsy (NB) is recommended prior to breast cancer surgery, the use of NB has been shown to vary according to patient socioeconomic status (SES), operating surgeon, and geographic region. We hypothesized that surgeons who work in the same peer referral network (defined by patient sharing) might have similar practice patterns with regard to NB, and that the magnitude of SES disparities might vary across networks. We therefore examined: 1) SES disparities in the receipt of NB, 2) variation in NB across networks, and 3) whether the association between SES and NB varied across networks.

Methods

We used the SEER database and 5% Medicare sample to examine all patients with a new diagnosis of breast cancer from 2004 through 2006. We used Medicare claims to construct peer groups of physicians based on patient-sharing ties. Patients were assigned to peer groups based on the surgeon who performed their definitive surgery. We defined a patient as having low SES if she was in the lowest quintile of area-level income. We used hierarchical generalized linear models (HGLM) to assess the association between low SES and receipt of NB, including random effects for the surgeon, peer group, and Hospital Referral Region (HRR). We then allowed the low SES effect to vary across peer groups in order to determine whether the association between SES and NB varied across groups.

Results

In the full sample of 14,552 patients, 9,498 (65%) received needle biopsy. In bivariable analysis, patients in the lowest income quintile were less likely to receive NB compared to all other patients (59% vs 67%, p&lt;.001). The majority of the variance (59%) in NB use was at the patient level, 22% was at the surgeon level, and 13.7% at the peer group level. The use of NB varied substantially across peer groups, with a median of 69% (interquartile range [51%, 84%]). Even after accounting for physician, peer group, and HRR variation, patients in the lowest stratum of SES were significantly less likely to have received NB compared to all other patient (OR = 0.88; p=.04). Finally, we found that the association between SES and NB varied significantly across referral networks (P&lt;0.05)

Conclusions

Patients with low SES are significantly less likely to receive NB prior to breast cancer surgery, and moreover the magnitude of this SES-related disparity varies significantly according to which referral networks are providing care. Future policies to increase NB rates and standardize care for all breast cancer patients may consider the implications of how care for patients with low SES varies across surgical provider networks.

Citation Format: Killelea BK, Herrin J, Soulos PR, Pollack CE, Forman HP, Yu J, Xu X, Tannenbaum S, Wang S, Gross CP. Socioeconomic disparities in needle biopsy prior to breast cancer surgery across physician referral networks [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-10-03.

Association between access to accelerated partial breast irradiation and use of adjuvant radiotherapy

BACKGROUND

The current study was performed to determine whether access to facilities performing accelerated partial breast irradiation (APBI) is associated with differences in the use of adjuvant radiotherapy (RT).

METHODS

Using the National Cancer Data Base, the authors performed a retrospective study of women aged ≥50 years who were diagnosed with early-stage breast cancer between 2004 and 2013 and treated with breast-conserving surgery (BCS). Facilities performing APBI in ≥10% of their eligible patients within a given year were defined as APBI facilities whereas those not performing APBI were defined as non-APBI facilities. All other facilities were excluded. The authors identified independent factors associated with RT use using multivariable logistic regression with clustering in the overall sample as well as in subsets of patients with standard-risk invasive cancer, low-risk invasive cancer, and ductal carcinoma in situ.

RESULTS

Among 222,544 patients, 76.6% underwent BCS plus RT and 23.4% underwent BCS alone. The likelihood of RT receipt in the overall sample did not appear to differ significantly between APBI and non-APBI facilities (adjusted odds ratio [AOR], 1.02; P = .61). Subgroup multivariable analysis demonstrated that among patients with standard-risk invasive cancer, there was no association between evaluation at an APBI facility and receipt of RT (AOR, 0.98; P = .69). However, patients with low-risk invasive cancer were found to be significantly more likely to receive RT (54.4% vs 59.5%; AOR, 1.22 [P<.001]), whereas patients with ductal carcinoma in situ were less likely to receive RT (56.9% vs 55.3%; AOR, 0.89 [P = .04]) at APBI facilities.

CONCLUSIONS

Patients who were eligible for observation were more likely to receive RT in APBI facilities but no difference was observed among patients with standard-risk invasive cancer who would most benefit from RT. Cancer 2017;123:502-511. © 2016 American Cancer Society.

Longer Periods Of Hospice Service Associated With Lower End-Of-Life Spending In Regions With High Expenditures

Hospice use is expected to decrease end-of-life expenditures, yet evidence for its financial impact remains inconclusive. One potential explanation is that the use of hospice may produce differential cost-savings effects by region because of geographic variation in end-of-life spending patterns. We examined 103,745 elderly Medicare fee-for-service beneficiaries in the Surveillance, Epidemiology, and End Results Program Medicare database who died from cancer in 2004-11. We created quintiles by the adjusted mean end-of-life expenditures per hospital referral region (HRR), and we examined HRR-level variation in the association between length of hospice service and expenditures across quintiles. Longer periods of hospice service were associated with decreased end-of-life expenditures for patients residing in regions with high average expenditures but not for those in regions with low average expenditures. Hospice use accounted for 8 percent of the expenditure variation between the highest and the lowest spending quintiles, which demonstrates the powers and limitations of hospice use for saving on costs.

Treatment Burden of Medicare Beneficiaries With Stage I Non-Small-Cell Lung Cancer

PURPOSE

To quantify the burden and complexity associated with treatment of Medicare beneficiaries with stage I non-small-cell lung cancer (NSCLC).

METHODS

Using the SEER-Medicare database, we conducted a retrospective cohort study of Medicare beneficiaries who were diagnosed with stage I NSCLC from 2007 to 2011 and who were treated with surgery, stereotactic body radiation therapy, or external beam radiation therapy. Main outcome measures were the number of days a patient was in contact with the health care system (encounter days), the number of physicians involved in a patient's care, and the number of medications prescribed. Logistic regression modeled the association between patient characteristics, treatment type, and high treatment burden (defined as ≥ 66 encounter days).

RESULTS

On average, 7,955 patients spent 1 in 3 days interacting with the health care system during the initial 60 days of treatment. Patients experienced a median of 44 encounter days with high variability (interquartile range [IQR], 29 to 66) in the 12 months after treatment initiation. The median number of physicians involved was 20 (IQR, 14 to 28), and the median number of medications prescribed was 12 (IQR, 8 to 17). Patients who were treated with surgery had high treatment burden (predicted probability, 21.6%; 95% CI, 20.2 to 23.1) compared with patients who were treated with stereotactic body radiation therapy (predicted probability, 16.1%; 95% CI, 12.9 to 19.3), whereas patients who were treated with external beam radiation therapy had the highest burden (predicted probability, 46.8%; 95% CI, 43.3 to 50.2).

CONCLUSION

The treatment burden imposed on patients with early-stage NSCLC was substantial in terms of the number of encounters, physicians involved, and medications prescribed. Because treatment burden varied markedly across patients and treatment types, future work should identify opportunities to understand and ameliorate this burden.

The Effect of Biologically Effective Dose and Radiation Treatment Schedule on Overall Survival in Stage I Non-Small Cell Lung Cancer Patients Treated With Stereotactic Body Radiation Therapy

PURPOSE

To determine the effect of biologically effective dose (BED₁₀) and radiation treatment schedule on overall survival (OS) in patients with early-stage non-small cell lung cancer (NSCLC) undergoing stereotactic body radiation therapy (SBRT).

METHODS AND MATERIALS

Using data from 65 treatment centers in the United States, we retrospectively reviewed the records of T1-2 N0 NSCLC patients undergoing SBRT alone from 2006 to 2014. Biologically relevant covariates, including dose per fraction, number of fractions, and time between fractions, were used to quantify BED₁₀ and radiation treatment schedule. The linear-quadratic equation was used to calculate BED₁₀ and to generate a dichotomous dose variable of <105 Gy versus ≥105 Gy BED₁₀. The primary outcome was OS. We used the Kaplan-Meier method, the log-rank test, and Cox proportional hazards regression with propensity score matching to determine whether prescription BED₁₀ was associated with OS.

RESULTS

We identified 747 patients who met inclusion criteria. The median BED₁₀ was 132 Gy, and 59 (7.7%) had consecutive-day fractions. Median follow-up was 41 months, and 452 patients (60.5%) had died by the conclusion of the study. The 581 patients receiving ≥105 Gy BED₁₀ had a median survival of 28 months, whereas the 166 patients receiving <105 Gy BED₁₀ had a median survival of 22 months (log-rank, P=.01). Radiation treatment schedule was not a significant predictor of OS on univariable analysis. After adjusting for T stage, sex, tumor histology, and Eastern Cooperative Oncology Group performance status, BED₁₀ ≥105 Gy versus <105 Gy remained significantly associated with improved OS (hazard ratio 0.78, 95% confidence interval 0.62-0.98, P=.03). Propensity score matching on imbalanced variables within high- and low-dose cohorts confirmed a survival benefit with higher prescription dose.

CONCLUSIONS

We found that dose escalation to 105 Gy BED₁₀ and beyond may improve survival in NSCLC patients treated with SBRT.

Surveillance Practice Patterns after Curative Intent Therapy for Stage I Non-Small-Cell Lung Cancer in the Medicare Population

OBJECTIVES

Recurrence after treatment for non-small cell lung cancer (NSCLC) is common, and routine imaging surveillance is recommended by evidence-based guidelines. Little is known about surveillance patterns after curative intent therapy for early stage NSCLC. We sought to understand recent practice patterns for surveillance of stage I NSCLC in the first two years after curative intent therapy in the Medicare population.

MATERIALS AND METHODS

Using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database we selected patients diagnosed with stage I NSCLC between 1998 and 2008. We studied adherence to surveillance guidelines based on specialty society recommendations for chest radiography and computed tomography (CT) scanning. We also tracked the use of Positron Emission Tomography (PET) scans, which are not recommended for surveillance. We calculated the percent of patients who received guideline-adherent surveillance imaging and used logistic regression to determine associations between patient and provider factors and guideline adherence.

RESULTS

Overall, 61.4% of patients received guideline-adherent surveillance during the initial 2 years after treatment. Use of CT scans in the first year after treatment increased from 47.4% in 1998-78.5% in 2008, and PET use increased from 5.8% to 28.9%. Adherence with surveillance imaging was associated with younger age, higher income, more comorbidities, access to primary care, and receipt of SBRT as the primary treatment.

CONCLUSIONS

Adherence to specialty society guidelines for surveillance after treatment for stage I NSCLC was poor in this population of Medicare beneficiaries, with less than two-thirds of patients receiving recommended imaging, and almost 30% receiving non-recommended PET scans.