How did the COVID crisis affect use of neoadjuvant therapy for patients with breast cancer?

e18708

Background: The COVID-19 pandemic has caused shifts in terms of cancer management, but the impact of this has not been well-elucidated in a contemporary cohort of patients in clinical practice in the US. We hypothesized that closure of operating rooms would increase the use of neoadjuvant therapy (NT) during the early pandemic period. Methods: The nationwide Flatiron Health database is a longitudinal electronic health record (EHR)-derived database, comprising de-identified, patient-level structured and unstructured data, curated via technology-enabled abstraction. These data originated from approximately 280 cancer clinics. We compared patients diagnosed with non-metastatic breast cancer during the early pandemic period (March 1 – June 30, 2020; group 1) with those diagnosed in the four month period prior (November 1, 2019 – February 29, 2020; group 2) and those diagnosed during the same period one year earlier (March 1 – June 30, 2019; group 3). Results: There were 174 patients in group 1, 277 in group 2, and 348 in group 3. Overall, 591 (74.1%) were ER/PR+HER2-, 100 (12.6%) were HER2+, and 106 (13.3%) were triple negative (TN). Patients in the three groups were equally likely to be ER/PR+HER2- (75.3% vs. 72.2% vs. 74.9%, p = 0.68), HER2+ (12.1% vs. 14.9% vs. 11%, p = 0.33), TN (12.6% vs. 12.7% vs. 14.2%, p = 0.83) and to be high risk by genomic testing (either Oncotype Dx or Mammaprint; p = 0.72). While there was no difference in the clinical stage (p = 0.36) nor patient age at diagnosis (p = 0.76) across the three groups, patients diagnosed during the early pandemic (group 1) were more likely to receive NT compared to those diagnosed one year earlier (group 3); 28.7% vs 16.4%, p < 0.01 (see table). The use of NT differed between the three groups in the ER/PR+her2- (p < 0.01) and her2+ patients (p = 0.05), but not in the TN patients (p = 0.61). There was no difference in the use of NT overall during the pandemic by geographic state (p = 0.32) nor practice setting (p = 0.23); NT was also similar by geographic state and practice setting when considering the ER/PR+HER2-, HER2+, and TNBC subsets. Conclusions: Despite similar clinicopathologic features as earlier time periods, there was an increased use of NT during the early pandemic when compared to the same period in the prior year. This was seen particularly in the ER/PR+HER2- group, suggesting an increased use of neoadjuvant endocrine therapy.[Table: see text]

Pathologic complete response (pCR) rates after neoadjuvant pertuzumab (P) and trastuzumab (H) administered concomitantly with weekly paclitaxel (T) and 5-fluorouracil/epirubicin/cyclophosphamide (FEC) chemotherapy for clinical stage I-III HER2-positive breast cancer

577

Background: Inclusion of H with chemotherapy has increased pathologic complete response (pCR) rates in HER2 positive breast cancer, and dual HER2 blockade involving H + P further increased efficacy. With dual HER2 blockade and taxane-based (+/-carboplatin followed by anthracycline) chemotherapies, pCR rates reach, 75% in estrogen receptor (ER) negative and 45% in ER+ patients. HER2 targeted therapies also increase the efficacy of anthracyclines but are not routinely combined due to potential cardiotoxicity. The goal of this phase II study was to assess pCR rate when H+P is administered during the entire treatment duration, including the anthracycline phase, of weekly T (80 mg/m2) x 12 followed by FE(75 mg/m2)C x 4 neoadjuvant chemotherapy. Methods: pCR (ypT0/is and ypN0) rate was assessed separately in ER+ and ER- cancers following Simon’s two-stage design to detect improvement in pCR rates to 90% and 70% in the ER- and ER+ cohorts, respectively. Eligibility included age <65, stage I-III, HER2+ disease, and normal cardiac function. Results: The ER- cohort completed full accrual of 25 patients: 23 completed therapy and surgery, 2 patients are still receiving treatment. The pCR rate is 78% (n=18, 95% CI:58-90%). The ER+ cohort was closed after 23 patients were accrued to the first stage due to lower than expected pCR of 26% (n=6, 95% CI:13-46%) at interim analysis. The incidence of grade 3/4 adverse events was 48% (n=24/50), the most common being neutropenia (n=12) and diarrhea (n=7). No patient experienced symptomatic congestive heart failure, one patient had a drop in LVEF to < 50% following completion of chemotherapy. Thirteen patients (27%) had a >10% asymptomatic drop in their LVEF but remained above 50%, LVEF returned to baseline by the next assessment in half of these cases. Conclusions: Neoadjuvant P and H administered concomitantly with weekly T followed by FEC resulted in 78% pCR rate in ER-/HER2+ cancers. This pCR rate is among the highest reported in the literature. The pCR rate was substantially lower in ER+ cancers. Clinical trial information: NCT01855828.

Safety of MEDI4736 (anti-PD-L1 antibody) administered concomitant with weekly nab-paclitaxel and dose dense doxorubicin/cyclophosphamide (ddAC) as neoadjuvant chemotherapy for stage I-III triple negative breast cancer (TNBC): A Phase I/II trial

572

Background: Pathologic complete response (pCR) rates to neoadjuvant chemotherapy in TNBC plateaued at 40% with existing regimens, the co-administration of an immune checkpoint inhibitor might increase pCR rate. The objective of the Phase I portion of this trial was to assess the safety of administering MEDI4736 concomitant with sequential taxane and anthracycline chemotherapy. Methods: The Phase I part followed the 3+3 design exploring two dose levels of MEDI4736 (3 and 10 mg/kg iv q2wk) in combination with weekly nab-paclitaxel (100 mg/m2) x 12 followed by ddAC x 4. Dose limiting toxicities (DLT) were evaluated during the entire 20 weeks of therapy and were defined as (1) gr 4 immune related adverse event (irAE), (2) gr 3 irAE that did not resolve to gr 2 within 3 days or to ≤ gr 1 within 14 days, (3) > gr 3 colitis or pneumonitis, (4) ≥ gr 3 non-irAE causally related to MEDI4736. Results: 3 patients completed therapy at the 3 mg/kg dose without any DLT, 1 additional patient refused further study medication because of recurrent gr 2 fatigue after 7 weeks of therapy. At the 10 mg/kg dose level, all 3 patients completed the nab-paclitaxel+MEDI4736 treatment without any DLT and 2 patients also completed 3 of the 4 planned treatments with ddAC without DLT. Among all 7 patients who started therapy, 1 at the 3 mg/kg group experienced gr 3 dehydration and dyspnea without chest X ray abnormalities which resolved within 48 hours with hydration. There were no other gr 3 AEs. Among the 3 patients who have completed therapy as per protocol (not including the patient who withdraw consent), 1 achieved pCR, 1 had minimal, and 1 had extensive residual cancer. No surgical AE were seen. All patients at the 10 mg/kg dose level will complete surgery by March 2017 and final Phase I toxicity and efficacy results will be presented. Conclusions: Concomitant administration of MEDI4736 10 mg/kg with weekly nab-paclitaxel and subsequently with ddAC neoadjuvant chemotherapy appears safe. The Phase II portion of the trial is open and will accrue a maximum of 50 patients to assess the efficacy of the combination. Clinical trial information: NCT02489448.

Abstract P1-09-01: African American women have lower pathologic complete response rates to neoadjuvant chemotherapy compared to white women for triple negative and HER 2 positive breast cancer

Introduction

Although racial disparities in breast cancer treatment have been well documented, data regarding differences in response to neoadjuvant chemotherapy are few. In 2010 the National Cancer Database (NCDB) included a new variable, documenting pathologic complete response (pCR) after neoadjuvant chemotherapy. The purpose of this study was to explore racial differences in the rates of pCR by molecular subtype.

Methods

The NCDB was queried to identify women diagnosed with invasive, stage 1-3 breast cancer in 2010 -2011 who received neoadjuvant chemotherapy. Univariate and multivariate logistic regression was performed to determine factors associated with likelihood of pCR.

Results

Out of 278,815 patients with known race and ethnicity, 27,300 (10%) received neoadjuvant chemotherapy. Of 17,970 where the outcome was known, 5,944 (33%) had a pCR. As seen in the table, there were no differences in response rate for ER/PR+ tumors, but compared to whites, non-hispanic black women had a lower rate of pCR for ER/PR- Her2+ and triple negative tumors. This difference persisted when adjusted for patient age, clinical T stage, clinical N stage, histology, grade, comorbidity index, facility type, geographic region, insurance status, and census-derived median income and education for the patient's zip code (OR 0.84, 95% CI: 0.77-0.93).

pCR rate by race and molecular subtype RacepCR number (%)p valueER/PR+, Her 2-Non-Hispanic White943/5129 (18.4%)reference Non-Hispanic Black204/1042 (19.6%)0.367 Non-Hispanic Asian/Pacific Islander59/291 (20.3%)0.420 Hispanic121/609 (19.9%)0.373ER/PR+, Her 2 +Non-Hispanic White852/2107 (40.4%)Reference Non-Hispanic Black143/380 (37.6%)0.304 Non-Hispanic Asian/Pacific Islander42/124 (33.9%)0.148 Hispanic92/224 (41.1%)0.854ER/PR-, Her 2 +Non-Hispanic White698/1295 (53.9%)Reference Non-Hispanic Black116/272 (42.6%)0.001 Non-Hispanic Asian/Pacific Islander66/112 (58.9%)0.306 Hispanic88/174 (50.6%)0.409ER/PR-, Her 2-Non-Hispanic White1318/3079 (42.8%)Reference Non-Hispanic Black416/1138 (36.6%)&lt;0.001 Non-Hispanic Asian/Pacific Islander64/165 (38.8%)0.310 Hispanic159/381 (41.7%)0.689

Conclusions

Non-hispanic black women have a lower likelihood of pCR after neoadjuvant chemotherapy compared to white women for triple negative and Her 2 positive breast cancer. It is unknown whether this is due to biologic differences in chemosensitivity or whether it represents treatment or socioeconomic differences that cannot be adjusted for in the current analysis.

Citation Format: Killelea BK, Chagpar AB, Horowitz NR, Pusztai L, Wang S, Mougalian S, Lannin DR. African American women have lower pathologic complete response rates to neoadjuvant chemotherapy compared to white women for triple negative and HER 2 positive breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P1-09-01.

Abstract P3-13-01: Impact of routine cavity shave margins on time and money: Results from the SHAVE trial

INTRODUCTION: Taking routine cavity shave margins (CSM) reduces positive margin and re-excision rates by 50%, but the impact of this technique on operative time and overall costs have not been well-elucidated.

METHODS: The SHAVE trial randomized 235 Stage 0-3 breast cancer patients undergoing partial mastectomy 1:1 to either have further cavity shave margins resected ("shave") or not ("no shave"). Randomization occurred intraoperatively after surgeons had completed standard partial mastectomy. Intraoperative time as well as actual direct costs incurred by the hospital were measured, for both the index case as well as any surgeries over the subsequent 90 days.

RESULTS: Median patient age was 61 (range; 33-94). 54 patients (23%) had invasive cancer, 45 (19%) had DCIS, and 125 (53%) had both. Median invasive tumor size was 1.1 cm (range; 0-6.5), and median DCIS size was 1.0 cm (range; 0-9.3). The "shave" and "no shave" groups were well-matched in terms of baseline characteristics, including the proportion having a sentinel node biopsy (75.6% vs. 69.8%, p=0.32) and/or axillary node dissection (9.2% vs. 7.8%, p=0.68) at the time of the initial surgery. The median number of additional CSM in the "shave" group was 4 (range; 3-6). At the initial surgery, those in the "shave" group had a longer operative time (median 76 vs. 66 minutes, p=0.005), and higher OR, pathology and total costs (see table). 48 patients required a subsequent surgery; 45 (93.8%) for margin clearance, 3 for sentinel lymph node biopsy alone (2 in the "shave" and 1 in the "no shave" group, p=1.00). There was a significantly lower re-excision rate for margins in the "shave" group (10.9% vs. 27.6%, p=0.001). Median time to re-excision was 22 days (range; 10-62). The mean cost of additional surgeries for those who required them was no different between the "shave" and "no shave" groups ($2636 vs. $3453, p=0.12); however, given the overall lower reoperation rate in the "shave" group (12.6% vs. 28.4%, p=0.003), the mean cost per patient for additional surgeries was significantly lower in the "shave" vs. "no shave" group. Taking into account all surgeries (including the index case and any additional surgeries within 90 days), there was no significant difference in cost (from a hospital perspective) between the two groups.

Mean (± SE) Costs per patient"Shave" (n=119)"No Shave" (n=116)p-valueIndex surgery:   OR costs$1315 (± $69)$1138 (± $52)0.042Pathology costs$1195 (± $43)$795 (± $48)&lt; 0.001Total costs$4758 (± $123)$4133 (± $119)&lt; 0.001Additional surgery:   OR costs$94 (± $24)$247 (± $44)0.003Pathology costs$51 (± $18)$112 (± $21)0.031Total costs$332 (± $88)$983 (± $189)0.002Total 90 day surgery costs:   OR costs$1409 (± $76)$1385 (± $64)0.808Pathology costs$1247 (± $49)$909 (± $52)&lt; 0.001Total costs$5090 (± $166)$5116 (± $214)0.925

CONCLUSIONS: Taking routine CSM is associated with increased time and cost for the index surgery, but this is offset by the cost savings of reduced re-excision rates. While the strategies of "shave" and "no shave" are similar in terms of 90 day hospital-related costs, taking CSM is associated with a lower need for reoperative surgery, thereby reducing patient angst and improving utilization of surgeon and OR time.

Citation Format: Chagpar AB, Longley PB, Horowitz NR, Killelea BK, Tsangaris TN, Li F, Butler M, Stavris K, Yao X, Harigopal M, Bossuyt V, Lannin DR, Pusztai L, Loftus M, Davidoff AJ, Gross CP. Impact of routine cavity shave margins on time and money: Results from the SHAVE trial. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-13-01.

Does three-dimensional intraoperative specimen imaging reduce the need for re-excision in breast cancer patients?

107

Background: Intraoperative specimen radiography is used by surgeons to evaluate partial mastectomy specimens to ensure that the lesion in question has been adequately removed. We sought to determine whether three-dimensional (3D) specimen imaging would better predict margin status and reduce the need for re-excision than conventional two-dimensional (2D) imaging. Methods: A prospective study using standard 2D as well as 3D specimen imaging was undertaken. Surgeons were asked whether the additional orthogonal view would change their management (i.e., result in further margins being taken intraoperatively), and the impact of this on final margin status. Results: Of the 100 women participating in the study, pathology results were available in 99. Of these, 10 had no residual tumor in the definitive specimen (either due to neoadjuvant chemotherapy, or due to the entire tumor being removed in the core biopsy). The remaining 89 patients formed the cohort of interest. 21 (23.6%) had DCIS, 18 (20.2%) had invasive cancer, and 50 (56.2%) had both. The median tumor size of the largest component was 1.7 cm (range; 0.2 – 8.1 cm). Based on the conventional two-dimensional imaging, surgeons stated they would take more tissue in 26 patients (29.2%). Of the 63 patients in whom no further excision would have been indicated on 2D imaging, the 3D imaging changed management in 4 patients (6.3%). Two of these patients would have had positive margins if the intraoperative resection done on the basis of the 3D imaging would have been omitted. Patients who surgeons felt, either by 2D or 3D intraoperative imaging, warranted intraoperative re-excision tended to have a closer initial margin than those in whom re-excision was thought not to be needed on the basis of intraoperative imaging (median 1.0 vs. 2.0 mm, p=0.038). Furthermore, patients in whom an immediate intraoperative margin was taken (either due to imaging or as a matter of routine) were less likely to require subsequent re-excision (13.4% vs. 40.9%, p=0.012). Conclusions: While 3D specimen imaging changes management in only 6.3% of cases, these data highlight the role of intraoperative immediate re-excision in potentially reducing re-excision rates.

Does removal of DCIS decrease the incidence of invasive breast cancer?

2

Background: SEER data have shown that as the incidence of high grade DCIS increases, high grade invasive cancer decreases; but this is not true for low-grade DCIS. The purpose of this study was first to determine what proportion of DCIS is calcified, and second to explore the relationship between grade, invasiveness, and calcification, in an effort to model the effect of removing mammographically detected DCIS. Methods: A single-institution, retrospective review was performed to identify all patients diagnosed with DCIS or invasive cancer from 2003-2011 who had corresponding mammograms. Results: There were 337 cases of pure DCIS, 309 cases of pure invasive tumor, and 847 cases of invasive tumor with a DCIS component. Grade 3 invasive cancers were much more likely to be associated with DCIS compared to grade 2 or 1 cancers. (83% vs. 70% vs. 54%, p<0.001). Of the 1,184 cases with DCIS or a DCIS component, 601 (51%) were associated with mammographic calcifications. The presence of calcifications was strongly associated with DCIS extent, comedo and micropapillary histology, presence of necrosis, and grade of DCIS (p<0.001 for each). High grade DCIS was almost twice as likely to be calcified compared to low grade. Extrapolation to SEER data suggests that there is a large reservoir of undetected, non-calcified DCIS and this is largely of low grade. Conclusions: These findings may explain why breast screening has reduced the incidence of high-grade invasive cancers, through diagnosing and excising high-grade DCIS, and yet has failed to reduce the incidence of low-grade cancers. [Table: see text]

The impact of survivorship care plans on knowledge among breast cancer survivors

124

Background: Despite the Institute of Medicine’s recommendation that all cancer survivors be provided with a survivorship care plan (SCP), there is limited information as to whether SCPs improve knowledge. The purpose of this study was to examine the impact of SCPs on patient knowledge regarding tumor stage, cancer treatment and potential risk of long-term side effects, surveillance recommendations, and perceived knowledge of their therapy. Methods: 75 English-speaking women over the age of 18 were enrolled in this prospective study. The participants’ treatment progress was tracked through the electronic medical record and used to create the SCP, which was given to them at the completion of treatment (defined as the time patients had completed radiation and/or chemotherapy treatments and initiated on hormonal therapy, if applicable). Knowledge of stage, treatments, potential side effects, and surveillance recommendations were assessed before receiving the SCP and again two months later. Accuracy of responses was compared using the McNemar test. Results: The median patient age was 56.8 ± 12.5 and 47.1% were stage I, 37.3% were stage II, and 15.7% were stage III. Participants were statistically more accurate in identifying their stage after receiving their SCP than at baseline (72.6% vs. 92.2%, p = 0.0016). While many patients were more accurate in the identification of the cancer treatment they received after SCP, the only significant improvement was in identifying 5-Fluorouracil as chemotherapy received (65.5% vs. 89.7%, p=0.0196). Patients were more accurate in identifying potential side effects but were only statistically more accurate at identifying leukemia as a risk factor (36.0% vs. 46.9%, p=0.0348). At baseline and follow up patients perceived that they had a high level of understanding as it related to their cancer stage (60.4%, and 66.7%), treatment (69.4%, and 71.4%), and surveillance recommendations (61.4%, and 54.6%). Conclusions: SCPs appear to improve patient knowledge in several important areas including basic and specific treatment details, surveillance recommendations, and potential side effects. Delivery of a SCPs is one strategy to improve knowledge of treatments received.

A pilot chemoprevention study of isopropanolic black cohosh extract in women with ductal carcinoma in situ

TPS1609

Background: Recent evidence suggests that black cohosh (Cimicifuga racemosa) may be a potential agent for breast cancer prevention. The active ingredients in black cohosh preparations appear to be triterpene glycosides. Recent preclinical data suggest several mechanisms by which triterpenes may prevent and treat breast cancer, including anti-proliferative, pro-apoptotic, anti-estrogenic, and anti-inflammatory effects. Epidemiologic data demonstrate a significant protective effect of black cohosh against the development of breast cancer in healthy women, and prolonged disease-free survival in breast cancer patients. There is also abundant evidence demonstrating the safety and tolerability of black cohosh in several clinical trials studying its use for treatment of hot flashes. We hypothesize that efficacy of black cohosh can be demonstrated in a pilot pre-operative window trial in a cohort of women with ductal carcinoma in situ (DCIS). Methods: In this trial, we treat women with a 2-5 week pre-operative course of commercial standardized isopropanolic black cohosh extract (20 mg orally twice per day). We aim to demonstrate a reduction in breast epithelial cell proliferation as measured by Ki-67 staining in regions of DCIS using traditional IHC staining and AQUA analysis. We also assess safety and tolerability of black cohosh through monitoring of patient adherence, liver function tests and serum hormone levels. 22 patients will be enrolled onto the trial. Sample size is based on power calculations for the specific study aim of determining the mean change in the levels of Ki67, using a targeted effect size. Assuming a 10% drop-out rate, a sample size of 20 patients will achieve 91% power to detect a 0.8 standard deviation of difference with a two-sided significance level at 0.05 using Wilcoxon signed-rank test. Eligible subjects are pre- and post-menopausal women ≥ 18 years of age newly diagnosed with DCIS histologically confirmed on breast core biopsy, prior to definitive excision. Women who have recently taken any agent known to affect Ki67 levels in the breast (e.g. hormone therapy) are excluded. Enrollment is currently ongoing with 10 of 22 patients accrued. Clinical trial registry number NCT01628536. Clinical trial information: NCT01628536.

SHAVE: A randomized controlled trial of routine shave margins versus standard partial mastectomy

TPS1145

Background: It is well known that partial mastectomy for breast cancer is associated with a positive margin rate of 20-40% in most series. This has led some surgeons to advocate for routine cavity shave margins as a means of reducing re-excision rates. Others, however, feel that such a practice may be unwarranted and question the volume of tissue removed, cosmetic outcome and increase in operative time. It is unclear which of these two approaches is optimal; therefore, a prospective randomized controlled trial was proposed. Methods: Given the primary endpoint of positive margin rates (defined as a margin of < 1mm), the study was powered to find a difference between 30% in the standard partial mastectomy group and 15% in the routine shave margin group. To reach a power of 80% with alpha of 5%, 122 patients were required in each arm; we therefore set N=250 with a 1:1 randomization scheme. Patients are evaluated preoperatively and all patients undergoing a partial mastectomy for stage 0-III breast cancer are eligible; including those who have completed neoadjuvant chemotherapy. Patients are stratified according to stage and randomized within strata. After informed consent, patients undergo a standard partial mastectomy (including specimen radiography as needed). Surgeons may resect additional tissue at that time according to their standard practice. At the completion of this procedure, the randomization envelope is opened in the operating room and surgeons are instructed to either shave (ie., take additional circumferential margins) or close (no shave). Patients will be followed for five years. Outcome measures include: positive margin/re-excision rate, local recurrence, volume of tissue resected, cosmetic outcome, and intraoperative time. To date, over 130 patients have accrued to this trial. Initial results are expected to be reported in 2014. Clinicaltrials.gov identifier: NCT01452399

Abstract P3-14-05: Recurrence in Patients Diagnosed with Ductal Carcinoma in Situ: Predictors and Prognostic Significance

Background: While patients diagnosed with ductal carcinoma in situ (DCIS) enjoy a favorable prognosis, recurrence after definitive management does occur in a subset of these patients. Factors influencing the development of recurrence, and the ultimate impact of this event on overall survival, remain poorly understood. We sought to determine clinicopathologic factors affecting recurrence and the prognostic role this plays in patients with DCIS.

Methods: A retrospective chart review of 205 consecutive patients who presented to an academic breast center with DCIS from 2000 to 2003 was conducted under an IRB-approved protocol. Non-parametric statistical analyses comparing patients who recurred to those who did not were then performed using SPSS Statistical Software.

Results: With a median follow-up of 8.5 years, 14 (6.8%) of the 205 patients who presented with DCIS between 2000–2003 had an invasive or in situ recurrence. The median age of all patients at the time of diagnosis of their initial DCIS was 55.5 years (range; 35.8–88.9). 51 (24.9%) had possible (albeit not definitive) microinvasion on their initial specimen. The majority were grade 2 (91; 44.4%), and 99 (48.3%) had comedo histology. The median size of DCIS on the initial excision was 10 mm (range; 0.1–80). All patients underwent definitive surgery to negative margins. The mean time to recurrence was 4.7 years (range; 1.1–10.6). On bivariate analysis, histologic grade of DCIS was the only factor that was significantly correlated with the risk of recurrence. Patients with grade 3 DCIS were more likely to develop a recurrence compared to those with grade 2 or grade 1 DCIS (12.3% vs. 3.3% vs. 0%, respectively, p = 0.032). Patient age, race, extent of DCIS, and histologic subtype were not associated with recurrence (see table below). 5-year actuarial overall survival was no different between those who developed a recurrence and those who did not (92.9 vs. 95.7%, p = 0.171).

Predictors of residual disease after breast-conserving surgery

168

Background: Locoregional failure after breast conserving surgery (BCS) is often due to undetected residual disease, and the risk of such residual disease frequently guides management. We sought to determine clinical and pathologic factors correlating with the finding of residual invasive cancer and/or DCIS in patients undergoing BCS. Methods: We performed a retrospective cohort study for all invasive and in situ breast cancer treated with BCS at a single institution in 2009. The main outcome variable of interest was residual disease determined by pathologic examinations of cavity shave margins or reexcision. Chart review and statistical analyses were performed to evaluate clinical and pathological factors correlating with residual DCIS or invasive cancer. Results: 256 in situ or invasive breast cancers were treated with BCS in 2009. Of these, 207 (80.9%) underwent additional resection either for close margins or as routine practice. These formed the cohort of interest for this study. 39 patients (18.8%) had residual DCIS and 22 (10.6%) had residual invasive disease. Age, race, histology, ER, PR, her-2-neu and margin distance for invasive disease did not predict the finding of residual DCIS nor invasive cancer. Lymphovascular invasion, while not predicting residual DCIS, was correlated with the finding of residual invasive disease (28.0% vs. 7.9%, p=0.007). Margin distance for DCIS was not predictive of residual invasive cancer but was predictive of residual DCIS. 33.8% of lesions with DCIS margins <1mm were associated with residual DCIS, while 3.4% of those with DCIS margins >5mm were associated with residual in situ disease (p=0.002). Increasing tumor size for invasive and in situ disease were associated with residual DCIS (median 19.5 vs. 13.0 mm, p=0.001 and 22.5 vs. 15.0 mm, p<0.001, respectively); however, neither size component was associated with residual invasive disease. Conclusions: While margin distance and tumor size are associated with residual DCIS in patients undergoing BCS, these are not correlated with residual invasive disease. Conversely, the finding of lymphovascular invasion predicts residual invasive cancer, but not DCIS. These factors may aid in risk stratification of patients and guide postoperative management.

Predictors of microinvasion and its prognostic role in ductal carcinoma in situ

188

Background: Predictors of microinvasion in ductal carcinoma in situ (DCIS) are not well understood. We sought to determine factors predicting microinvasion and the prognostic role this plays in patients with DCIS. Methods: A retrospective cohort study of 205 consecutive patients presenting to the Yale Breast Center with DCIS from 2000 through 2003 was performed. A chart review was conducted and bivariate and multivariate analyses comparing patients with and without possible microinvasion were performed. Statistical analyses were done using SPSS software version 19. Results: Of the 205 patients who presented with DCIS and were treated with surgical excision, 51 (24.9%) had evidence of possible microinvasion on final pathology. The median age of all patients was 53.0 years (range 35.8 to 88.9). On bivariate analysis, patients with microinvasion had larger areas of DCIS, and were more likely to have high grade DCIS, of solid type, associated with necrosis and microcalcifications (see table). There was a trend towards white women having a higher rate of microinvasion than black women (26.9% vs. 8.7%, p=0.061). On multivariate analysis, none of these factors were independent predictors of microinvasion. With a median follow-up of 8.5 years, there was no difference in the likelihood of recurrence in the microinvasion vs. no microinvasion groups (6.0% vs. 7.2%, p=1.000). 5-year actuarial overall survival was also not different between the two groups (96% vs. 94%, p=0.202, respectively). Conclusions: Patients with larger DCIS size, higher grade, solid histology, necrosis, and microcalcifications have a higher likelihood of microinvasion. However, the presence of possible microinvasion does not significantly increase risk of recurrence or decrease survival. [Table: see text]

Accuracy of breast MRI in predicting pathologic tumor size

1109

Background: MRI use as a preoperative planning tool is increasing in women with breast cancer, yet the correlation between MRI and pathologic size of cancers is unclear. The purpose of this study was to determine the accuracy of MRI in predicting pathologic tumor size, and factors that affect this correlation. Methods: Clinicopathologic and imaging data from 84 patients diagnosed with invasive or in situ breast cancer from September 2010 to October 2011 who had preoperative MRI were reviewed. 12 patients who had neoadjuvant chemotherapy were excluded. MRI detected 147 lesions in the remaining 72 patients. Concordance between MRI and pathology size was determined using Spearman rho coefficients, and factors affecting the accuracy of MRI in predicting tumor size within +/- 0.5 cm were determined. Results: There was a modest correlation between MRI and pathology size for all MRI detected lesions (benign or malignant) with a Spearman coefficient of 0.53. Of the 147 MRI detected lesions, 45 (30.6%) had pathologic and MRI size correlating within +/- 0.5 cm; 76 (51.7%) were overestimated (>0.5cm) by MRI, and 26 (17.7%) were underestimated (>0.5cm). 101 (68.7%) of the 147 lesions were found to be malignant (either with invasive disease or DCIS). In this subgroup, 35 lesions (34.7%) had an MRI size within +/- 0.5 cm of the pathologic size; 40 (39.6%) were overestimated by MRI and 26 (25.7%) were underestimated. Patient age, tumor histology, LVI and grade did not predict concordance between pathologic and MRI size. However, small MRI lesion size more accurately correlated with pathologic tumor size. While 51.1% of tumors that had concordant MRI and pathologic findings within 0.5 cm were <1 cm on MRI, no tumor found to be > 5 cm on MRI was within +/-0.5 cm on final pathology (p=0.001). Conclusions: MRI accurately predicts pathologic tumor size only when the size of the lesion on MRI is <1 cm.

Splenic arteriogastric fistula from plication of a gastric ulcer. Therapeutic embolization

Massive upper gastrointestinal hemorrhage occurred in a 66-year-old man 18 days after suture plication of a giant gastric ulcer. Arteriography revealed a splenic arteriogastric fistula. Therapeutic embolization with tissue adhesive (bucrylate) successfully controlled hemorrhage after failure of attempted particulate embolization.