Development of a Multilevel Model to Identify Patients at Risk for Delay in Starting Cancer Treatment.

Importance: Delays in starting cancer treatment disproportionately affect vulnerable populations and can influence patients' experience and outcomes. Machine learning algorithms incorporating electronic health record (EHR) data and neighborhood-level social determinants of health (SDOH) measures may identify at-risk patients. Objective: To develop and validate a machine learning model for estimating the probability of a treatment delay using multilevel data sources. Design, Setting, and Participants: This cohort study evaluated 4 different machine learning approaches for estimating the likelihood of a treatment delay greater than 60 days (group least absolute shrinkage and selection operator [LASSO], bayesian additive regression tree, gradient boosting, and random forest). Criteria for selecting between approaches were discrimination, calibration, and interpretability/simplicity. The multilevel data set included clinical, demographic, and neighborhood-level census data derived from the EHR, cancer registry, and American Community Survey. Patients with invasive breast, lung, colorectal, bladder, or kidney cancer diagnosed from 2013 to 2019 and treated at a comprehensive cancer center were included. Data analysis was performed from January 2022 to June 2023. Exposures: Variables included demographics, cancer characteristics, comorbidities, laboratory values, imaging orders, and neighborhood variables. Main Outcomes and Measures: The outcome estimated by machine learning models was likelihood of a delay greater than 60 days between cancer diagnosis and treatment initiation. The primary metric used to evaluate model performance was area under the receiver operating characteristic curve (AUC-ROC). Results: A total of 6409 patients were included (mean [SD] age, 62.8 [12.5] years; 4321 [67.4%] female; 2576 [40.2%] with breast cancer, 1738 [27.1%] with lung cancer, and 1059 [16.5%] with kidney cancer). A total of 1621 (25.3%) experienced a delay greater than 60 days. The selected group LASSO model had an AUC-ROC of 0.713 (95% CI, 0.679-0.745). Lower likelihood of delay was seen with diagnosis at the treating institution; first malignant neoplasm; Asian or Pacific Islander or White race; private insurance; and lacking comorbidities. Greater likelihood of delay was seen at the extremes of neighborhood deprivation. Model performance (AUC-ROC) was lower in Black patients, patients with race and ethnicity other than non-Hispanic White, and those living in the most disadvantaged neighborhoods. Though the model selected neighborhood SDOH variables as contributing variables, performance was similar when fit with and without these variables. Conclusions and Relevance: In this cohort study, a machine learning model incorporating EHR and SDOH data was able to estimate the likelihood of delays in starting cancer therapy. Future work should focus on additional ways to incorporate SDOH data to improve model performance, particularly in vulnerable populations.

Adverse Events Reported by Patients With Cancer After Administration of a 2-Dose mRNA COVID-19 Vaccine.

BACKGROUND: Most safety and efficacy trials of the SARS-CoV-2 vaccines excluded patients with cancer, yet these patients are more likely than healthy individuals to contract SARS-CoV-2 and more likely to become seriously ill after infection. Our objective was to record short-term adverse reactions to the COVID-19 vaccine in patients with cancer, to compare the magnitude and duration of these reactions with those of patients without cancer, and to determine whether adverse reactions are related to active cancer therapy. PATIENTS AND METHODS: A prospective, single-institution observational study was performed at an NCI-designated Comprehensive Cancer Center. All study participants received 2 doses of the Pfizer BNT162b2 vaccine separated by approximately 3 weeks. A report of adverse reactions to dose 1 of the vaccine was completed upon return to the clinic for dose 2. Participants completed an identical survey either online or by telephone 2 weeks after the second vaccine dose. RESULTS: The cohort of 1,753 patients included 67.5% who had a history of cancer and 12.0% who were receiving active cancer treatment. Local pain at the injection site was the most frequently reported symptom for all respondents and did not distinguish patients with cancer from those without cancer after either dose 1 (39.3% vs 43.9%; P=.07) or dose 2 (42.5% vs 40.3%; P=.45). Among patients with cancer, those receiving active treatment were less likely to report pain at the injection site after dose 1 compared with those not receiving active treatment (30.0% vs 41.4%; P=.002). The onset and duration of adverse events was otherwise unrelated to active cancer treatment. CONCLUSIONS: When patients with cancer were compared with those without cancer, few differences in reported adverse events were noted. Active cancer treatment had little impact on adverse event profiles.

Early Prostate-Specific Antigen Kinetics for Low- and Intermediate-Risk Prostate Cancer Treated With Definitive Radiation Therapy.

PURPOSE: This study used a patient-specific model to characterize and compare ideal prostate-specific antigen (PSA) kinetics for low- and intermediate-risk prostate cancer after definitive radiation treatment with conventionally fractionated, hypofractionated, stereotactic body radiation therapy, or brachytherapy, both high-dose and low-dose rate. METHODS AND MATERIALS: This retrospective analysis includes low- and intermediate-risk patients with prostate cancer treated between 1998 and 2018 at an National Cancer Institute-designated comprehensive cancer center. Demographics and treatment characteristics were prospectively collected. Patients had at least 2 PSA measurements within 24 months of treatment and were free from biochemical recurrence. The incidence of, time to, and risk factors for PSA nadir (nPSA) and bounce (bPSA) were analyzed at 24 months after radiation therapy. Ideal PSA kinetics were characterized for each modality and compared. RESULTS: Of 1042 patients, 45% had low-risk cancer, 37% favorable intermediate risk, and 19% unfavorable intermediate risk. nPSAs were higher for ablative modalities, both as absolute nPSA and relative to initial PSA. Median time to nPSA ranged from 14.8 to 17.1 months. Over 50% treated with nonablative therapy (conventionally fractionated, hypofractionated, and low-dose rate) reached an nPSA threshold of ≤0.5 ng/mL compared with 23% of stereotactic body radiation therapy and 33% of high-dose rate cohorts. The incidence of bPSA was 13.3% and not affected by treatment modality, Gleason score, or prostate volume. PSA decay rate was faster for ablative therapies in the 6- to 24-month period. CONCLUSIONS: Analysis of PSA within 24 months after radiation therapy revealed ablative therapies are associated with a latent PSA response and higher nPSA. Multivariable logistics modeling revealed younger age, initial PSA above the median, presence of bPSA, and ablative therapy as predictors for not achieving nPSA ≤0.5 ng/mL. PSA decay rate appears to be faster in ablative therapies after a latent period. Understanding the different PSA kinetic profiles is necessary to assess treatment response and survey for disease recurrence.

Ten-Year Update of a Randomized, Prospective Trial of Conventional Fractionated Versus Moderate Hypofractionated Radiation Therapy for Localized Prostate Cancer.

PURPOSE: The previously published single institution randomized prospective trial failed to show superiority in the 5-year biochemical and/or clinical disease failure (BCDF) rate with moderate hypofractionated intensity-modulated radiation therapy (H-IMRT) versus conventionally fractionated IMRT (C-IMRT). We now present 10-year disease outcomes using updated risk groups and definitions of biochemical failure. METHODS: Men with protocol-defined intermediate- and high-risk prostate adenocarcinoma were randomly assigned to receive C-IMRT (76 Gy in 38 fractions) or H-IMRT (70.2 Gy in 26 fractions). Men with high-risk disease were all prescribed 24 months of androgen deprivation therapy (ADT) and had lymph node irradiation. Men with intermediate risk were prescribed 4 months of ADT at the discretion of the treating physician. The primary endpoint was cumulative incidence of BCDF. We compared disease outcomes and overall mortality by treatment arm, with sensitivity analyses for National Comprehensive Cancer Network (NCCN) risk group adjustment. RESULTS: Overall, 303 assessable men were randomly assigned to C-IMRT or H-IMRT. The median follow-up was 122.9 months. Per updated NCCN risk classification, there were 28 patients (9.2%) with low-risk, 189 (62.4%) with intermediate-risk, and 86 (28.4%) with high-risk prostate cancer. The arms were equally balanced for clinicopathologic factors, except that there were more black patients in the C-IMRT arm (17.8% v 7.3%; P = .02). There was no difference in ADT use (P = .56). The 10-year cumulative incidence of BCDF was 25.9% in the C-IMRT arm and was 30.6% in the H-IMRT arm (hazard ratio, 1.31; 95% CI, 0.82 to 2.11). The two arms also had similar cumulative 10-year rates of biochemical failure, prostate cancer-specific mortality, and overall mortality; however, the 10-year cumulative incidence of distant metastases was higher in the H-IMRT arm (rate difference, 7.8%; 95% CI, 0.7% to 15.1%). CONCLUSION: H-IMRT failed to demonstrate superiority compared with C-IMRT in long-term disease outcomes.

Treatment Trends and Outcomes for Patients With Lymph Node-Positive Cancer of the Penis.

Importance: Penile cancer is an uncommon disease with minimal level I evidence to guide therapy. The National Comprehensive Cancer Network (NCCN) guidelines advocate a lymph node dissection (LND) or radiotherapy with consideration of perioperative chemotherapy for all patients with lymph node-positive (LN+) penile cancer without metastasis. Objectives: To determine temporal trends in use of chemotherapy for patients with LN+ penile cancer without metastasis and to evaluate outcomes between those who did or did not receive LND, chemotherapy, and radiotherapy. Design, Setting, and Participants: The US National Cancer Database (NCDB) was queried for all 1123 patients with LN+, squamous cell carcinoma of the penis without metastasis from January 1, 2004, through December 31, 2014. Temporal trends were assessed using Cochran-Armitage tests. Multivariable logistic models were used to examine the association between treatments, clinicopathologic variables, and receipt of chemotherapy. Kaplan-Meier analyses with log-rank tests and multivariable Cox regressions were used to analyze overall survival. Data were analyzed between January 2017 and September 2017. Main Outcomes and Measures: Use of chemotherapy over time. Survival outcomes by receipt or nonreceipt of LND, radiotherapy, and chemotherapy. Results: Of 1123 patients identified, most were white (924 [82.3%]) vs African American (141 [12.6%]) or of other or unknown race (58 [5.2%]). The age of most patients (727 [64.7%]) was between 50 and 75 years, and 750 patients (66.8%) underwent an LND. From 2004 to 2014, the use of systemic therapy significantly increased (26 of 68 patients, 38.2% vs 65 of 136, 47.8%; P < .001). However, only 177 of 335 patients with N3 disease (52.8%) received chemotherapy (N1: 106 of 338, 31.4%; N2: 178 of 450, 39.6%). Following adjustment, older patients (>76 years: OR, 0.28; 95% CI, 0.15-0.50; P < .001) were less likely to receive chemotherapy. Patients who received radiotherapy (OR, 4.38; 95% CI, 3.10-6.18; P < .001) and those patients with N2 (OR, 1.62; 95% CI, 1.16-2.27; P = .005) or N3 (OR, 2.32; 95% CI, 1.67-3.22; P < .001) cancer were more likely to receive chemotherapy. On multivariable analysis, LND (HR, 0.64; 95% CI, 0.52-0.78; P < .001) was associated with better overall survival, while neither chemotherapy (HR, 1.01; 95% CI, 0.80-1.26; P = .95) nor radiotherapy (HR, 0.85; 95% CI, 0.70-1.04; P = .11) was associated with overall survival. Conclusions and Relevance: In hospitals reporting to the NCDB, only 66.8% of patients with LN+ penile cancer received an LND. While chemotherapy use has increased since 2004, rates remain low (52.8% for patients with N3 cancer). Receipt of LND, but not chemotherapy or radiotherapy, is associated with overall survival. This may reflect the aggressive natural history of penile cancer as well as the inherent analysis limitation of a relatively small sample size. These data highlight opportunities to improve adherence to guideline-recommended care.

Differences in Survival Associated with Performance of Lymph Node Dissection in Patients with Invasive Penile Cancer: Results from the National Cancer Database.

PURPOSE: Inguinal lymphadenectomy remains under performed in patients with invasive penile cancer. Using a large national cancer registry we assessed temporal trends in inguinal lymphadenectomy performance and evaluated the impact of the procedure on survival in patients in whom inguinal lymphadenectomy was an absolute indication (T1b-4 N0/x-1) according to NCCN® (National Comprehensive Cancer Network®) Guidelines®. MATERIALS AND METHODS: We queried the National Cancer Database for all cases of nonmetastatic, T1b-4 N0/x-1 squamous cell carcinoma of the penis from 2004 to 2014. Multivariable logistic regression models adjusting for patient, demographic, and clinicopathological characteristics were used to examine the association between available covariates and receipt of inguinal lymphadenectomy. Cox proportional hazards regression analysis was then done to assess the impact of clinical and pathological variables on overall survival. Propensity score weighted analysis was performed to assess the effect of inguinal lymphadenectomy on overall survival. RESULTS: A total of 2,224 patients met analysis criteria, of whom 606 (27.2%) underwent inguinal lymphadenectomy. Following adjustment the procedure was more likely in younger patients, those who presented with palpable adenopathy (cN1), those treated at an academic facility and those with a more contemporary diagnosis. On survival analysis controlling for all known and measured confounders inguinal lymphadenectomy was associated with improved overall survival (HR 0.79, 95% CI 0.74-0.84, p <0.001). CONCLUSIONS: At hospitals that report to the National Cancer Database the overall rate of inguinal lymphadenectomy in patients with invasive penile cancer was only 27.2%. Inguinal lymphadenectomy was associated with increased overall survival, justifying the procedure as an important quality metric for performance reporting in patients with invasive penile cancer.