Strengths and Limitations of Registries in Surgical Oncology Research

Use of the National Cancer Database in Identifying Disparities in Cancer Treatment

The National Cancer Database (NCDB) is a large data set of cancer patients treated in the United States. In recent decades, many papers have utilized this database to study disparities in cancer care and outcomes. In this review, we present the strengths and limitations of using the NCDB to study cancer disparities, which should be taken into account by researchers who use the NCDB in this setting.

Adjusting for Population Differences in the National Cancer Database to Better Represent United States Cancer Cases: A Reference Tool for Researchers

BACKGROUND

The National Cancer Database (NCDB) is widely used in US cancer outcomes research, but its reliance on Commission on Cancer-approved hospitals can underrepresent certain populations, skew data, and limit generalizability of findings. Current literature is representative up through 2014. We sought to adjust NCDB cancer cases to better reflect total US cancer population in a useful way for cancer outcomes research.

METHODS

Incident cancer cases in the NCDB from 2016-2020 were compared with the US Cancer Statistics (USCS) database, which contains nearly 100% of new cancer cases. NCDB case coverage was defined as percentage of cases the NCDB represents of USCS cases. Coverage was determined for the entire cohort (age 20+ years), and sub-analyses were performed for age, sex, race/ethnicity, residence location, and cancer sites.

RESULTS

From 2016-2020, 6,515,675 cancer cases were diagnosed in the NCDB and 9,311,593 in the USCS, yielding 70% NCDB case coverage over 5 years, which increased from 68 to 73%. The lowest case coverage was among men, 85+-year-olds, American Indian/Alaskan Native people, and Hispanic/Latino individuals (65%, 59%, 42%, and 55%). The Mountain region was the least represented (49%) as was nonmetropolitan residence (64%). Similar underrepresentation was seen among top cancers. Missingness of data was also captured.

CONCLUSIONS

Though NCDB's representation of US cancer cases is improving, gaps remain, including age, sex, race/ethnicity, and residence location, further exacerbated by missing variables. We provide investigators using the NCDB with a way to represent cancer case data to better tailor research questions and frame outcomes.

Machine learning models for prediction of lymph node metastasis in patients with gastric cancer: a Chinese single-centre study with external validation in an Asian American population

OBJECTIVE

To develop and validate machine learning (ML)-based models to predict lymph node metastasis (LNM) in patients with gastric cancer (GC).

DESIGN

Retrospective cohort study.

SETTING

Second Affiliated Hospital of Soochow University.

PARTICIPANTS

A total of 500 inpatients from the Second Affiliated Hospital of Soochow University, collected retrospectively between 1 April 2018 and 31 March 2023, were used as the training set, while 824 Asian patients from the Surveillance, Epidemiology and End Results database comprised the external validation set.

MAIN OUTCOME MEASURES

Prediction models were developed using multiple ML algorithms, including logistic regression, support vector machine, k-nearest neighbours, naive Bayes, decision tree (DT), gradient boosting DT, random forest and artificial neural network (ANN). The predictive value of these models was validated and evaluated through receiver operating characteristic curves, precision-recall (PR) curves, calibration curves, decision curve analysis and accuracy metrics.

RESULTS

Among the ML algorithms, the ANN outperformed others, achieving the highest accuracy (0.722; 95% CI: 0.692 to 0.751), precision (0.732; 95% CI: 0.694 to 0.776), F1 score (0.733; 95% CI: 0.695 to 0.773), specificity (0.728; 95% CI: 0.684 to 0.770) and area under the PR curve (0.781; 95% CI: 0.740 to 0.821) in the external validation results. Moreover, it demonstrated superior calibration and clinical utility. Shapley Additive Explanations analysis identified the depth of invasion, tumour size and Lauren classification as the most influential predictors of LNM in patients with GC. Furthermore, a user-friendly web application was developed to provide individual prediction results.

CONCLUSIONS

This study introduces an accurate, reliable and clinically applicable approach for predicting the risk of LNM in patients with GC. The model demonstrates its potential to enhance the personalised management of GC in diverse populations, supported by external validation and an accessible web application for practical use.

Travel distance to tertiary sarcoma centres does not influence oncological presentation or outcomes

Aims

Soft-tissue sarcomas (STSs) are rare cancers with centralized care advocated to consolidate resources and expertise. However, geographical challenges, particularly in countries like Canada, can increase travel distances for patients. The impact of travel distance on sarcoma presentation and outcomes remains unclear, particularly in single-payer healthcare systems with centralized care.

Methods

A retrospective cohort analysis was conducted on 1,570 patients with STS who underwent surgical resection at a Canadian tertiary referral centre between January 2010 and January 2021. Patients were divided into those living ≤ 50 km and > 50 km from the centre. Demographics, tumour characteristics, treatment methods, and survival outcomes were analyzed. A Cox regression model was constructed to evaluate predictors of overall survival.

Results

Patients living > 50 km from the centre (n = 700) travelled a mean of 176 km (SD 250), while those ≤ 50 km (n = 870) travelled a mean of 24.8 km (SD 13.8). There were no significant differences in disease presentation, time to definitive treatment, use of systemic therapies, or functional outcomes between the two groups. The two-year and five-year overall survival rates were similar between the groups (83.1% (95% CI 80.1% to 86.1%) vs 83.8% (95% CI 81.8% to 85.8%) and 72.1% (95% CI 69.1% to 75.1%) vs 72.5% (95% CI 69.5% to 75.5%), respectively). The regression model demonstrated that age, higher tumour grade, depth, and lower income were predictive of worse overall survival, while distance travelled was not an independent predictor of survival.

Conclusion

Contrary to previous studies, our findings suggest that travel distance did not influence disease presentation or survival outcomes in STS patients treated at a centralized sarcoma centre. This challenges previous notions regarding the impact of travel distance on cancer outcomes, and supports the effectiveness of centralized care models, even in geographically vast regions.

Metastatic Breast Cancer: Prolonging Life in Routine Oncology Care

Overall survival (OS) of patients with metastatic breast cancer (MBC) has improved within controlled clinical trials. Whether these advances translate into improved OS in routine care is controversial. We therefore analyzed retrospectively unselected female patients from five oncology group practices and one university outpatient clinic, whose initial diagnosis of MBC was between 1995 and 2022. A total of 1610 patients with a median age of 63 years (23-100) were evaluated. In all, 82.9% had hormone-receptor-positive disease, and 23.8% were HER2-positive. Evaluation in time cohorts by initial MBC diagnosis date showed a continuous prolongation of median OS from 31.6 months (0.5-237.3+) (1995-2000) to 48.4 months (0.4-61.1+) (2018-2022) (p = 0.003). Univariable analyses showed a significant dependence on the time cohort of diagnosis, metastatic status at initial diagnosis, age at metastasis, hormone and HER2 status, general condition, metastasis localization, and the number of affected organs. A multivariable analysis revealed a significant dependence of survival probability on receptor status, general condition, and number of metastatic sites, as well as the time between initial breast cancer diagnosis and the diagnosis date of MBC in months. In sum, OS of patients with MBC has improved continuously and significantly in routine care over the last 27 years.

Sublobar resection is associated with less lymph nodes examined and lower delivery of adjuvant therapy in patients with 1.5- to 2.0-cm clinical IA2 non-small-cell lung cancer: a retrospective cohort study

OBJECTIVES

CALGB140503, in which nodal sampling was mandated, reported non-inferior disease-free survival for patients undergoing sublobar resection (SLR) compared to lobectomy (L). Outside of trial settings, the adequacy of lymphadenectomy during SLR has been questioned. We sought to evaluate whether SLR is associated with suboptimal lymphadenectomy, differences in pathologic upstaging and survival in patients with 1.5- to 2.0-cm tumours using real-world data.

MATERIALS AND METHODS

Using the National Cancer Database(2018-2019), we evaluated patients with 1.5- to 2.0-cm non-small-cell lung cancer who underwent resection (sublobar versus lobectomy). We studied factors associated with nodal upstaging (logistic regression) and survival (Cox regression and Kaplan-Meier method) after propensity matching to adjust for differences among groups.

RESULTS

Among 3196 patients included, SLR was performed in 839 (26.3%) (of which 588 were wedge resections) and L was performed in 2357 (73.7%) patients. More patients undergoing SLR (21.7%) compared to L (2.1%) had no lymph nodes sampled (P < 0.001). Those undergoing SLR had fewer total lymph nodes examined (4 vs 11, P < 0.001) and were less likely to have pathologic nodal metastases (4.7% vs 9%, P < 0.001) compared to L. Multivariable analysis identified L [adjusted odds ratio (aOR) 2.21, 95% confidence interval, 1.47-3.35] to be independently associated with pathologic N+ disease. Overall survival was not associated with the type of procedure but was significantly decreased in those with N+ disease.

CONCLUSIONS

Despite comparable overall survival to L, SLR is associated with suboptimal lymphadenectomy in patients with 1.5-2.0 cm non-small-cell lung cancer. Surgeons should be careful to perform adequate lymphadenectomy when performing SLR to mitigate nodal under-staging and to identify appropriate patients for systemic therapy.

Underutilization of Systemic Therapy in Patients With NSCLC Undergoing Pneumonectomy: A Missed Opportunity for Survival

Introduction

Recent trials have reported promising results with the addition of immunotherapy to chemotherapy for patients with locally advanced NSCLC, but in practice, the proportion of patients who receive systemic therapy (ST) has historically been low. Underutilization of ST may be particularly apparent in patients undergoing pneumonectomy, in whom the physiologic insult and surgical complications may preclude adjuvant therapy (ADJ). We, therefore, evaluated the use of ST for patients with NSCLC undergoing pneumonectomy.

Methods

We queried the National Cancer Database, including all patients with NSCLC who underwent pneumonectomy between 2006 and 2018. Logistic regression was used to identify associations with ST and neo-ADJ (NEO). Overall survival was compared after propensity score matching (1:1) patients undergoing ST to those undergoing surgery alone using Kaplan-Meier and Cox regression methods.

Results

A total of 2619 patients were identified. Among these, 12% received NEO, 43% received ADJ, and 45% surgery alone. Age younger than 65 years (adjusted odds ratio [aOR] = 1.53, 95% confidence interval; [CI]: 1.10-2.11), Asian ethnicity (aOR = 2.68, 95% CI: 1.37-5.23), treatment at a high-volume center (aOR = 1.39, 95% CI: 1.06-1.81), and private insurance (aOR = 1.42, 95% CI: 1.05-1.94) were associated with NEO, whereas age younger than 65 years (aOR = 1.95, 95% CI: 1.61-2.38), comorbidity index less than or equal to 1 (aOR = 1.66, 95% CI: 1.29-2.16), and private insurance (aOR = 1.47, 95% CI: 1.20-1.80) were associated with any ST. In the matched cohort, ST was associated with better survival than surgery (adjusted hazard ratio = 0.67, 95% CI: 0.58-0.78).

Conclusions

A high proportion of patients who undergo pneumonectomy do not receive ST. Patient and socioeconomic factors are associated with the receipt of ST. Given its survival benefit, emphasis should be placed on multimodal treatment strategies, perhaps with greater consideration given to neoadjuvant approaches.

Patterns of Care for Breast Radiotherapy in Italy: Breast IRRadiATA (Italian Repository of Radiotherapy dATA) Feasibility Study

Aim. Breast IRRADIATA (Italian Repository of RADIotherapy dATA) is a collaborative nationwide project supported by the Italian Society of Radiotherapy and Clinical Oncology (AIRO) and the Italian League Against Cancer (LILT). It focuses on breast cancer (BC) patients treated with radiotherapy (RT) and was developed to create a national registry and define the patterns of care in Italy. A dedicated tool for data collection was created and pilot tested. The results of this feasibility study are reported here. Methods. To validate the applicability of a user-friendly data collection tool, a feasibility study involving 17 Italian Radiation Oncology Centers was conducted from July to October 2021, generating a data repository of 335 BC patients treated between January and March 2020, with a minimum follow-up time of 6 months. A snapshot of the clinical presentation, treatment modalities and radiotherapy toxicity in these patients was obtained. A Data Entry Survey and a Satisfaction Questionnaire were also sent to all participants. Results. All institutions completed the pilot study. Regarding the Data Entry survey, all questions achieved 100% of responses and no participant reported spending more than 10 min time for either the first data entry or for the updating of follow-up. Results from the Satisfaction Questionnaire revealed that the project was described as excellent by 14 centers (82.3%) and good by 3 (17.7%). Conclusion. Current knowledge for the treatment of high-prevalence diseases, such as BC, has evolved toward patient-centered medicine, evidence-based care and real-world evidence (RWE), which means evidence obtained from real-world data (RWD). To this aim, Breast IRRADIATA was developed as a simple tool to probe the current pattern of RT care in Italy. The pilot feasibility of IRRADIATA encourages a larger application of this tool nationwide and opens the way to the assessment of the pattern of care radiotherapy directed to other cancers.