Faculty Bibliography

AI is now a cornerstone of modern dataset analysis. In many real world applications, practitioners are concerned with controlling specific kinds of errors, rather than minimizing the overall number of errors. For example, biomedical screening assays may primarily be concerned with mitigating the number of false positives rather than false negatives. Quantifying uncertainty in AI-based predictions, and in particular those controlling specific kinds of errors, remains theoretically and practically challenging. We develop a strategy called multidimensional informed generalized hypothesis testing (MIGHT) which we prove accurately quantifies uncertainty and confidence given sufficient data, and concomitantly controls for particular error types. Our key insight was that it is possible to integrate canonical cross-validation and parametric calibration procedures within a nonparametric ensemble method. Simulations demonstrate that while typical AI based-approaches cannot be trusted to obtain the truth, MIGHT can be. We apply MIGHT to answer an open question in liquid biopsies using circulating cell-free DNA (ccfDNA) in individuals with or without cancer: Which biomarkers, or combinations thereof, can we trust? Performance estimates produced by MIGHT on ccfDNA data have coefficients of variation that are often orders of magnitude lower than other state of the art algorithms such as support vector machines, random forests, and Transformers, while often also achieving higher sensitivity. We find that combinations of variable sets often decrease rather than increase sensitivity over the optimal single variable set because some variable sets add more noise than signal. This work demonstrates the importance of quantifying uncertainty and confidence-with theoretical guarantees-for the interpretation of real-world data.

BACKGROUND:Pancreatic adenosquamous carcinoma has historically poor overall survival, and the impact of perioperative chemotherapy remains unclear. We aimed to evaluate the impact of various chemotherapy regimens in patients with resected adenosquamous carcinoma. METHODS:Patients with resected adenosquamous carcinoma were identified from 3 high-volume programs between 2001 and 2022. We analyzed their clinicopathologic data and used Kaplan-Meier survival curves to assess the median overall survival and recurrence-free survival with 95% confidence intervals. Prognostic factors were assessed with a multivariable Cox-regression model adjusting for resectability status and Clavien-Dindo complications. RESULTS:Among 168 patients, cohorts of neoadjuvant chemotherapy (41, 24%) and upfront surgery (127, 76%) showed similar demographics and TNM staging. The median overall survival was shorter in the neoadjuvant chemotherapy cohort compared with the upfront surgery cohort (13 vs 21 months, P = .133). Median overall survival by treatment approach was no chemotherapy (4 months), only neoadjuvant chemotherapy (8 months), only adjuvant therapy (24 months), and both neoadjuvant chemotherapy and adjuvant therapy (17 months). Recurrence-free survival data (69 patients) showed upfront surgery had significantly longer recurrence-free survival compared with neoadjuvant chemotherapy (18 months vs 5 months, P = .046). Multivariable analysis showed adjuvant therapy was associated with improved overall survival (hazard ratio, 0.27; P < .001), whereas age ≥65 (hazard ratio, 1.79, P = .030) was associated with worse overall survival. CONCLUSION/CONCLUSIONS:The outcomes of resected adenosquamous carcinoma remain poor. Patients receiving neoadjuvant chemotherapy exhibited shorter recurrence-free survival and median overall survival, suggesting minimal benefit of neoadjuvant chemotherapy in treating this aggressive cancer. Meanwhile, adjuvant therapy appears to be protective but requires further investigation.

BACKGROUND:Intraductal papillary mucinous neoplasm (IPMN)-derived pancreatic cancer was previously categorized into tubular, colloid, and oncocytic subtypes. Intraductal oncocytic papillary neoplasms (IOPN) has long been associated with superior prognosis/indolent behavior, however, there is discordant emerging evidence. This study aimed to investigate this conflicting literature. METHODS:Patients with resected IOPN-derived and IPMN-derived pancreatic cancer were identified from six international centers. Log-rank tests compared time to (TtR) and survival after (SAR) recurrence and five-year overall survival (OS). A multivariable mixed model was used to determine hazard ratios (HR) with confidence intervals (95%CI) for five-year survival. RESULTS:Of 879 patients, 20 (2%) had IOPN-derived pancreatic cancer. Most patients had T1 (55%) or N0 (70%) disease. IOPN and colloid IPMN-derived pancreatic cancers had similar recurrence rates (25% vs. 24%), while recurrence was more common in tubular IPMN-derived pancreatic cancer (42%, p < 0.001). IOPN-derived pancreatic cancer displayed a longer TtR and SAR compared to colloid and tubular IPMN-derived pancreatic cancers. IOPN-derived and colloid IPMN-derived cancers demonstrated significantly lower 5-year mortality risks compared to tubular IPMN-derived cancers (74% and 27% risk reduction, respectively; p < 0.05). CONCLUSION/CONCLUSIONS:IOPN-derived pancreatic cancers have excellent OS. However, some patients have poor prognostic factors and are at risk for both local and systemic recurrence. Given more indolent disease progression given delayed TtR and prolonged SAR compared to colloid and tubular IPMN-derived pancreatic cancers, there may be a role for prolonged surveillance.

BACKGROUND:Evidence regarding the optimal surgical approach for pancreatic neck/body cancer with portomesenteric vein (PV) involvement is scarce. We aimed to clarify the current practice using an international survey. METHODS:An online survey was distributed to members of nine international associations and study groups. Surgeons who performed pancreatectomy with PV resection (PVR) in the last 12 months were asked about three clinical scenarios with different PV involvement: scenarios A (<90°; length 1 cm), B (<90°; length 3 cm), and C (90-180°; length 3 cm), with or without common hepatic artery (CHA) involvement. PVR was defined according to the ISGPS definition. RESULTS:Overall, 222 surgeons from 49 countries in 6 continents completed the survey. The most selected procedures were left pancreatectomy with PVR ISGPS-type 1 for scenario A (52.3 %), PVR ISGPS-type 2 for B (28.8 %), and pancreatoduodenectomy with PVR ISGPS-type 3 for C (28.4 %). In patients with CHA involvement, the most selected procedures were left pancreatectomy without arterial reconstruction for A (57.7 %) and B (50.0 %), and total pancreatectomy for C (29.7 %). CONCLUSIONS:The survey illustrates the heterogeneity in surgical management of pancreatic neck/body cancer with PV involvement, indicating the need for prospective studies and guidelines.

BACKGROUND:Antithrombotic therapy (AT) aims to strike a balance between preventing thromboembolic and hemorrhagic complications. However, evidence for AT management after pancreatectomy with vascular reconstruction is lacking. We aimed to provide an overview of the current use of AT for pancreatic surgery with vascular reconstructions. PATIENTS AND METHODS/METHODS:A web-based survey was distributed to 123 surgeons from high-volume pancreas centers (>50 pancreatic resections/year). AT management after different types of vascular reconstruction were investigated. An "aggressive" protocol was defined as the use of any AT protocol other than prophylactic heparin, aspirin, or their combination. RESULTS:The survey was completed by 80 surgeons (59% Europe, 30% USA, 11% Asia). In Europe/Asia, prophylactic heparin was the most commonly reported protocol after partial venous resection/end-to-end anastomosis/human graft (71%/65%/50%, respectively), and an "aggressive" protocol (86%) was the most frequently used after prosthetic graft reconstruction. Conversely, in the USA, prophylactic heparin + aspirin was the most commonly reported protocol after all types of venous reconstruction. Following arterial reconstruction, heparin + aspirin was the most commonly reported protocol, regardless of region. An "aggressive" protocol was more frequently used in Europe/Asia (odds ratio (OR) 1.28; p < 0.001) and following vein reconstruction with either human graft (OR 1.2; p = 0.007) or prosthetic graft (OR 1.56, p <0.001), while ultrasound (OR 1.65; p < 0.001) and arterial reconstruction (OR 1.64; p < 0.001) were significantly associated with antiplatelet use. CONCLUSIONS:In an international cohort of high-volume pancreas surgeons, significant variation in the use of AT following pancreatectomy with vascular reconstruction was observed. This variation was driven by geographical differences and the type of vascular reconstructions performed. In an international cohort of high-volume pancreas surgeons, this Worldwide Snapshot Survey analyzed the current use of antithrombotic therapy for pancreatic surgery with vascular reconstruction. A significant heterogeneity in antithrombotic practice was found and it was mainly driven by geographical differences and the type of vascular reconstructions performed.

Patients with pancreatic ductal adenocarcinoma (PDAC) frequently present with liver metastases, which severely limit treatment options and prognosis. In other cancers, treatment of liver disease can improve outcomes and similar approaches are being explored in PDAC. Clinical data for locoregional control of pancreatic cancer liver metastases (PCLM) are limited, and histotripsy offers a new noninvasive tool for disease control. This study evaluates the preliminary safety, efficacy, and imaging findings of histotripsy in patients with PCLM.

BACKGROUND:Neoadjuvant therapy (NAT) for pancreatic ductal adenocarcinoma (PDAC) is increasingly being used. The aim of this study was to evaluate the association between type, duration, and sequencing of adjuvant therapy (AT) after NAT and overall survival (OS) in patients with resected PDAC. METHODS:Patients receiving NAT and resection for PDAC (2010-2019) at two high-volume pancreatic surgery centers were included and stratified into groups on the basis of NAT regimen: gemcitabine-based NAT, 5-fluorouracil (5FU)-based NAT, or switched NAT regimen. The maximally selected rank statistic was used to determine the optimal NAT duration. Univariate and multivariable Cox proportional hazards models were used to assess the association between NAT regimen and OS, and between AT and OS. RESULTS:Of 651 patients, 200 (30.7%) received gemcitabine-based NAT, 362 (56%) received 5FU-based NAT, and 89 (13.7%) switched NAT regimen. Median OS in patients receiving gemcitabine-based NAT was 19 months (95% confidence interval (CI) 17-25 months), compared with 26 months (95% CI 24-31 months) in patients receiving 5FU-based NAT (hazard ratio (HR) 0.81, 95% CI 0.66-0.99, p = 0.04) and 21 months (95% CI 16-26 months) in patients who switched NAT regimen (HR 0.98, 95% CI 0.73-1.29, p = 0.86). Optimal NAT duration was 3.6 months in the complete cohort. Receiving AT was associated with improved survival (HR 0.61, 95% CI 0.43-0.86, p < 0.001), but its association was reduced after a NAT duration of ≥5 months (adjuvant chemotherapy × NAT duration ≥ 5 months: HR 1.50, 95% CI 1.00-2.24, p = 0.048). CONCLUSIONS:Patients receiving 5FU-based NAT showed improved survival compared with patients receiving gemcitabine-based NAT before surgery for PDAC. Adjuvant therapy improved survival, particularly in patients with shorter NAT duration.