Dana-Farber Brigham Cancer Center

The integration of generative AI (GenAI) and large language models (LLMs) in healthcare presents both unprecedented opportunities and challenges, necessitating innovative regulatory approaches. In this perspective, we discuss the risks of GenAI and LLM-based medical devices, the limitations of current medical device regulation frameworks when applied to GenAI or LLMs, and advocate for global collaboration in regulatory science research through engaging multidisciplinary expertise and focusing on the needs of diverse populations.

PALB2 plays a crucial role in DNA repair by interacting with BRCA1 and BRCA2. To clarify the clinicopathological impact of somatic PALB2 variants, we analyzed tumor DNA extracted from formalin-fixed, paraffin-embedded (FFPE) breast cancer specimens from 81 patients. Targeted next-generation sequencing identified 1 514 variants, including 84 pathogenic or likely pathogenic variants and 9 genes with high mutation frequency; 2 deleterious somatic PALB2 variants were detected in 33% of samples. To explore transcriptional effects, we analyzed publicly available METABRIC mRNA expression data and found that elevated PALB2 mRNA levels were significantly associated with ER-/PR- tumors and poorer survival outcomes. High PALB2 expression remained an independent predictor of shorter overall survival and relapse-free survival. These findings highlight the clinical importance of somatic PALB2 mutations and the prognostic value of PALB2 expression in breast cancer.

Active surveillance (AS) is the preferred management approach for patients with low-risk prostate cancer (PC); yet whether younger patients with favorable-intermediate-risk (FIR) PC experience increased mortality-risk when electing AS remains unknown. We evaluated all-cause, PC-specific, and non-PC-specific mortality (ACM, PCSM, and non-PCSM) in younger patients with FIR PC managed with either AS/watchful-waiting (WW) or immediate definitive treatment, stratified by race. We conducted a retrospective cohort study using SEER data (2010-2020). Patients included were < 60 years-old with FIR PC. The primary outcome was ACM, secondary outcomes PCSM and non-PCSM. Multivariable Cox and Fine-Gray competing-risk regressions were used, adjusting for known prognostic factors. Interaction by race (White vs underrepresented minority [URM]) was explored. Statistical significance was set at p < 0.025 (Bonferroni-adjusted). Among 3,832 patients, 127 died (3.31%), including 18 of the 127 deaths from PC (14.17%). Initial treatment with RP/RT did not significantly reduce ACM or non-PCSM compared to AS/WW in White (ACM AHR, 0.92; 95% CI, 0.44-1.94; non-PCSM AHR, 1.36; 95% CI, 0.53-3.46) or URM patients (ACM AHR, 0.68; 95% CI, 0.33-1.43; non-PCSM AHR, 1.04; 95% CI, 0.44-2.44). However, after adjustment for multiplicity RP/RT significantly reduced PCSM-risk compared to AS/WW in URM (AHR, 0.03; 95% CI, 0.00-0.48; p = 0.01), but not in White patients (AHR, 0.21; 95% CI, 0.05-0.88; p = 0.03) although the median follow-up was 6.5-months longer in URM patients undergoing AS/WW compared to RP/RT. Early mortality-risks were similar and low in patients age < 60 years with FIR PC managed with AS/WW compared to RP/RT, irrespective of race.

Racial disparities in cancer care are well-documented.Although a variety of interventions have been implemented to address these disparities, the extent to which they have narrowed gaps in treatment remains an area of active research.In this context, Lynch et al 1 examined whether previously observed racial disparities in treatment of early-stage lung cancer persist in more contemporary practice.They used the Surveillance, Epidemiology, and End Results Program-Medicare database from 2005 to 2019. 1 Their primary outcome was receipt of curative therapy, defined as either surgical resection or radiotherapy, with a secondary outcome of receiving preferred treatment, defined as either lobectomy or stereotactic body radiation therapy (SBRT).Using multivariable analysis, the authors adjusted for race, year of diagnosis, a race-year interaction term, variables reflecting access to primary care, and additional clinical and sociodemographic characteristics.They found that non-Hispanic Black patients with early-stage lung cancer were less likely than their non-Hispanic White counterparts to undergo curative therapy, largely owing to lower rates of surgical resection. 1Black patients were also less likely to undergo SBRT during its early adoption (2011-2013); this disparity narrowed and was no longer statistically significant in the 2017 to 2019 study period.Overall, rates of radiotherapy were not consistently different by race.These associations persisted after adjustment for clinical and sociodemographic factors, highlighting the need for additional progress before achieving racial equity in cancer care.Taken together, this study 1 suggests that the largest and most persistent inequity is not in the type of radiation or surgery delivered once a modality is chosen, but in whether Black patients access surgical treatment at all, and whether they shared equitably in SBRT adoption in earlier years.These findings of racial disparities in the rate of surgical resection, with no consistent racial differences in radiotherapy receipt, invite comparison between the 2 modalities.Upstream factors are multifaceted, with patient factors such as comorbidities and frailty and system factors such as referral patterns and support for navigating treatment, all likely contributing.Regarding patient factors, the authors found that, on average, Black patients had higher rates of medical comorbidities and were less likely to have a primary care physician. 1ack patients who are less connected to primary care may be referred later, less frequently, or not at all for surgical evaluation.However, access to a primary care physician alone may not fully mitigate these difficulties.Recent work showed differential specialty referral patterns based on race. 2 These disparate referral patterns may negatively affect a patient's surgical candidacy, with downstream consequences for the likelihood of receiving curative oncologic treatment.Relatedly, social determinants of health such as financial toxicity or health literacy may play a role in influencing adherence to curative treatment for lung cancer, highlighting the need for financial navigation and assistance with transportation and other nonmedical costs to reduce financial toxicity and address cost-related barriers that undermine adherence to curative treatment. 3nch et al 1 also utilized a claims-based frailty measure that was validated with patient-reported outcomes.This approach may partially mitigate, but not fully eliminate, the influence of clinical bias in preoperative decision-making.Recent work demonstrating that racially and ethnically minoritized patients are rated as more frail on subjective than objective measures would suggest that subjective assessments could disproportionately disqualify Black patients from surgery, even when claimsbased frailty appears similar.Therefore, broader use of objective frailty measures in both practice and research is an important step toward equitable cancer care.