MSI score by clinically determined MSI status in 718 study participants

Supplementary Table 4 shows list of recurrently mutated genes in sporadic colorectal cancer

Supplementary Table 2 shows detailed characteristics of the study cohorts

Supplementary Table 6 shows estimated average genetic ancestry in 718 colorectal cancer patients by study cohort

Supplementary Table 14 shows associations between global genetic ancestry and tumor mutational signatures in 718 colorectal cancer patients

Mutational landscape in 128 Latino colorectal cancer patients

Mutational pattern of top mutated genes in 718 colorectal cancer patients

Estimated genetic ancestry for 718 colorectal cancer patients stratified by study cohort

ABSTRACT Purpose To compare patient outcomes across oncology hospitalist (OH) and general hospitalist (GH) teams and understand related team practices. Methods Using an explanatory sequential mixed‐methods design, we compared unplanned inpatient admissions to the GH or OH teams at our academic medical center between 2018 and 2022 using propensity‐score matching to balance characteristics across teams. Primary outcomes included outpatient oncology follow‐up, 30‐day readmission, and discharge to hospice. Outcomes were modeled with multivariable logistic regression, adjusted for age, race, and comorbidity. We conducted semi‐structured interviews with OH (n = 6), GH (n = 7), and oncology consult team (n = 4) clinicians. Themes were identified through rapid qualitative analysis and consensus discussions between two qualitative researchers. Results The matched sample included 1082 GH and 361 OH team admissions. Median length of stay was 5 days (IQR 3,8) for GH and 6 days for OH (IQR 4, 10) (p < 0.01). The OH team had higher adjusted odds of discharge to hospice (aOR: 1.9, 95% CI: 1.1–3.5), outpatient oncology follow‐up within 30 days (aOR: 4.2, 95% CI: 3.1–5.6), but also 30‐day readmissions (aOR: 1.8, 95% CI: 1.3–2.5) compared to the GH team. The OH team communicated directly with outpatient oncology throughout hospitalizations and sent discharge summaries with follow‐up recommendations. In contrast, the GH team relied on the consulting oncology team to communicate with outpatient oncologists and had variable follow‐up practices. Conclusion Engagement with outpatient oncologists throughout patient admissions may contribute to more effective discharge planning. Inpatient teams can use this strategy to help improve hospital transitions of care for patients with cancer.

Importance:Studying the molecular profiles of mixed-type and branch-duct (BD) intraductal papillary mucinous neoplasms (IPMNs) is important to understand the underlying biological basis for higher malignant potential of mixed-type IPMNs. Objective:To compare mutation patterns in mixed-type vs BD-IPMNs through cyst fluid next-generation sequencing (NGS) analysis using PancreaSeq NGS. Design, Setting, and Participants:For this cohort study, pancreatic cyst fluid specimens from 31 medical centers were sent to a centralized NGS lab for analysis between January 2018 and February 2020. Patients with IPMNs (based on KRAS and GNAS mutant status) and with available main pancreatic duct (MPD) size data were included. Patients with main-duct IPMNs and with MPD 10 mm or greater were excluded. Mixed-type IPMNs were defined as IPMNs with an MPD of 5 to 9 mm, and BD-IPMNs were defined as IPMNs with an MPD less than 5 mm. High-risk mutations (HRMs) were categorized as alterations in TP53, SMAD4, CTNNB1, and mTOR genes. Advanced neoplasia was defined as IPMNs with invasive carcinoma or high-grade dysplasia. Data were analyzed from June 1 to 5, 2025. Main Outcomes and Measures:Primary outcomes included rates of HRMs and the co-occurrence of 2 or more HRMs in mixed-type IPMNs and BD-IPMNs. Results:Among 674 patients with IPMNs, 202 had mixed-type IPMN, and 472 had BD-IPMN. There were 379 female patients (56.2%) and 295 male patients (43.8%); the mean (SD) age was 70.3 (9.6) years. HRMs were observed in 106 patients (16%), with TP53, SMAD4, and MTOR mutations more common in mixed-type IPMNs. Of the 674 patients, 167 patients underwent surgical resection, and these 167 patients had final surgical pathology available. Overall, mixed-type IPMNs had significantly higher rates of HRMs (62 [31%] vs 44 [9.3%]; P < .001) and co-occurrence of 2 or more HRMs (25 [12.4%] vs 14 [3%]; P < .001) compared with BD-IPMNs. On multivariate logistic regression, mixed-type IPMNs were independently associated with HRMs (odds ratio, 3.42; 95% CI, 1.72-6.82). Preoperative NGS detection of HRMs showed 90% sensitivity, 100% specificity, a positive predictive value (PPV) of 100%, and a negative predictive value (NPV) of 86% for predicting advanced neoplasia in mixed-type IPMNs. The presence of any worrisome feature or high-risk stigmata showed a sensitivity of 100%, very low specificity of 13.3%, PPV of 67%, and NPV of 100%. Conclusions and Relevance:This study found that mixed-type IPMNs are more likely to harbor HRMs associated with advanced neoplasia. Cyst fluid NGS is highly sensitive and specific for predicting advanced neoplasia in patients with mixed-type IPMNs and should therefore be considered to help upgrade or downgrade risk based on the presence or absence of other worrisome features.

Supplementary Table 10 shows association between global genetic ancestry and individual somatic mutations in 718 colorectal cancer patients

Supplementary Table 5 shows estimated average genetic ancestry proportions in 718 colorectal cancer patients by ethnicity

Supplementary Table 3 shows summary of the reference populations from the 1KGP and PAGE databases

Supplementary Table 1 shows distribution of demographic and clinical characteristics among 718 colorectal cancer patients

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive and devastating disease, with most patients facing fatality due to resistance to chemotherapy. The chemotherapeutic cocktail FOLFIRINOX serves as a first line of therapy, and its success can be predicted by the activity of carboxylesterase 2 (CES2), an enzyme responsible for metabolically activating irinotecan (CPT-11), a component of FOLFIRINOX, where low CES2 activity forecasts resistance. Such predictive power enables clinicians to decide whether FOLFIRNOX is worth pursuing prior to administration. Despite this, current methods to measure CES2 activity are insufficient, and a feasible and translatable method for clinical use remains to be developed. To this end, we use Benz-AP, a CES2 ratiometric fluorescent sensor, to detect CES2 activity toward predicting FOLFIRINOX sensitivity. Leveraging this sensor, we demonstrate strong correlations between Benz-AP readouts of CES2 activity and CES2 expression (R2 = 0.9812, P = 0.0047), CPT-11 response (R2 = 0.9181, P = 0.0417), and FOLFIRINOX sensitivity (R2 = 0.9813, P = 0.0417) in PDAC cell lines. These correlations were not observed with TOPO-1 expression levels, SN-38, or FOLFOX treatment. Likewise, strong and significant correlations between Benz-AP measurements and sensitivity of CES2-activated drugs (R2 = 0.7633, P = 0.0156) were also observed in a panel of primary patient PDAC cell lines. Furthermore, we demonstrate the compatibility of Benz-AP for use in ex vivo biopsy sampling by fine needle aspirates, where we detected a significant difference (P < 0.05) in CES2 activity across samples. Collectively, these results demonstrate the applicability of Benz-AP for predicting FOLFIRINOX prognosis across patient samples ranging in biological complexity and advocate for the translation of such technologies in the clinic to guide clinician choice of chemotherapeutic treatment.

Supplementary Table 12 shows association between TMB or dMMR/MSI status and each 10% increase in global genetic ancestry in 718 colorectal cancer patients

Mutation spectrum of APC, TP53 and KRAS in 718 colorectal cancer patients

Supplementary Table 7 shows commonly mutated genes in 718 colorectal cancer patients

Supplementary Table 9 shows association between ethnicity and gene mutation status of recurrently mutated genes in 597 colorectal cancer patients