A Computer Vision Algorithm to Predict Superior Mesenteric Artery Margin Status for Patients with Pancreatic Ductal Adenocarcinoma.

OBJECTIVE: To evaluate the feasibility of developing a computer vision algorithm that uses preoperative computed tomography (CT) scans to predict superior mesenteric artery (SMA) margin status in patients undergoing Whipple for pancreatic ductal adenocarcinoma (PDAC), and to compare algorithm performance to that of expert abdominal radiologists and surgical oncologists. SUMMARY BACKGROUND DATA: Complete surgical resection is the only chance to achieve a cure for PDAC; however, current modalities to predict vascular invasion have limited accuracy. METHODS: Adult patients with PDAC who underwent Whipple and had preoperative contrast-enhanced CT scans were included (2010-2022). The SMA was manually annotated on the CT scans, and we trained a U-Net algorithm for SMA segmentation and a ResNet50 algorithm for predicting SMA margin status. Radiologists and surgeons reviewed the scans in a blinded fashion. SMA margin status per pathology reports was the reference. RESULTS: Two hundred patients were included. Forty patients (20%) had a positive SMA margin. For the segmentation task, the U-Net model achieved a Dice Similarity Coefficient of 0.90. For the classification task, all readers demonstrated limited sensitivity, although the algorithm had the highest sensitivity at 0.43 (versus 0.23 and 0.36 for the radiologists and surgeons, respectively). Specificity was universally excellent, with the radiologist and algorithm demonstrating the highest specificity at 0.94. Finally, the accuracy of the algorithm was 0.85 versus 0.80 and 0.76 for the radiologists and surgeons, respectively. CONCLUSIONS: We demonstrated the feasibility of developing a computer vision algorithm to predict SMA margin status using preoperative CT scans, highlighting its potential to augment the prediction of vascular involvement.

Outpatient and Ambulatory Extended Recovery Robotic Hepatectomy: Multinational Study of 307 Cases.

BACKGROUND: For open minor hepatectomy, morbidity and recovery are dominated by the incision. The robotic approach may transform this "incision dominant procedure" into a safe outpatient procedure. STUDY DESIGN: We audited outpatient (less than 2 midnights) robotic hepatectomy at 6 hepatobiliary centers in 2 nations to test the hypothesis that the robotic approach can be a safe and effective short-stay procedure. Establishing early recovery after surgery programs were active at all sites, and home digital monitoring was available at 1 of the institutions. RESULTS: A total of 307 outpatient (26 same-day and 281 next-day discharge) robotic hepatectomies were identified (2013 to 2023). Most were minor hepatectomies (194 single segments, 90 bi-segmentectomies, 14 three segments, and 8 four segments). Thirty-nine (13%) were for benign histology, whereas 268 were for cancer (33 hepatocellular carcinoma, 27 biliary, and 208 metastatic disease). Patient characteristics were a median age of 60 years (18 to 93 years), 55% male, and a median BMI of 26 kg/m 2 (14 to 63 kg/m 2 ). Thirty (10%) patients had cirrhosis. One hundred eighty-seven (61%) had previous abdominal operation. Median operative time was 163 minutes (30 to 433 minutes), with a median blood loss of 50 mL (10 to 900 mL). There were no deaths and 6 complications (2%): 2 wound infections, 1 failure to thrive, and 3 perihepatic abscesses. Readmission was required in 5 (1.6%) patients. Of the 268 malignancy cases, 25 (9%) were R1 resections. Of the 128 with superior segment resections (segments 7, 8, 4A, 2, and 1), there were 12 positive margins (9%) and 2 readmissions for abscess. CONCLUSIONS: Outpatient robotic hepatectomy in well-selected cases is safe (0 mortality, 2% complication, and 1.6% readmission), including resection in the superior or posterior portions of the liver that is challenging with nonarticulating laparoscopic instruments.

Impact of safety-net hospital burden on achievement of textbook oncologic outcomes following resection in for stage I-IV colorectal cancer.

BACKGROUND AND OBJECTIVES: Textbook oncologic outcome (TOO) is a benchmark for high-quality surgical cancer care but has not been studied at safety-net hospitals (SNH). The study sought to understand how SNH burden affects TOO achievement in colorectal cancer. METHODS: The National Cancer Database was queried for colorectal cancer patients who underwent resection for stage I-III plus stage IV with liver-only metastases (2010-2019). TOO was defined as R0 resection, AJCC-compliant lymphadenectomy (>12 nodes), no prolonged LOS, no 30-day mortality/readmission, and receipt of stage-appropriate adjuvant chemotherapy. RESULTS: Of 487,195 patients, 66.7% achieved TOO. Lower achievement was explained by adequate lymphadenectomy (87.3%), non-prolonged LOS (76.3%), and receipt of adjuvant chemotherapy in stage III (60.3%) and IV (54.1%). Treatment at high burden hospitals (HBH, >10% Medicaid/uninsured) was a predictor of non-TOO (Stage I/II: OR 0.83, III: OR 0.86, IV: OR 0.83; all p < 0.001). Achieving TOO was associated with decreased mortality (Stage I/II: HR 0.49, III: HR 0.48, IV: HR 0.57; all p < 0.001), and HBH treatment was a predictor of mortality (Stage I/II: HR 1.09, III: HR 1.05, IV: HR 1.07; all p < 0.05). CONCLUSIONS: Treatment at higher SNH burden hospitals was associated with less frequent TOO achievement and increased mortality. Quality improvement targets include receipt of adjuvant chemotherapy and avoidance of prolonged LOS.

Surgical Cancer Care in Safety-Net Hospitals: a Systematic Review.

BACKGROUND: Tertiary medical centers in the USA provide specialized, high-volume surgical cancer care, contributing standards for quality and outcomes. For the most vulnerable populations, safety-net hospitals (SNHs) remain the predominant provider of both complex and routine healthcare needs. The objective of this study was to evaluate access to and quality of surgical oncology care within SNHs. METHODS: A comprehensive and systematic review of the literature was conducted using PubMed, EMBASE, and Cochrane Library databases to identify all studies (January 2000-October 2021) reporting the delivery of surgical cancer care at SNHs in the USA (PROSPERO #CRD42021290092). These studies describe the process and/or outcomes of surgical care for gastrointestinal, hepatopancreatobiliary, or breast cancer patients seeking treatment at SNHs. RESULTS: Of 3753 records, 37 studies met the inclusion criteria. Surgical care for breast cancer (43%) was the most represented, followed by colorectal (30%) and hepatopancreatobiliary (16%) cancers. Financial constraints, cultural and language barriers, and limitations to insurance coverage were cited as common reasons for disparities in care within SNHs. Advanced disease at presentation was common among cancer patients seeking care at SNHs (range, 24-61% of patients). Though reports comparing cancer survival between SNHs and non-SNHs were few, results were mixed, underscoring the variability in care seen across SNHs. CONCLUSIONS: These findings highlight barriers in care facing many cancer patients. Continued efforts should address improving both access and quality of care for SNH patients. Future models include a transition away from a two-tiered system of resourced and under-resourced hospitals toward an integrated cancer system.

Paradoxical Association of Hyperglycemia and Surgical Complications Among Patients With and Without Diabetes.

Importance: Surgical complications associated with perioperative hyperglycemia are conventionally associated with diabetes, but, paradoxically, prior cohort studies have found that patients without diabetes have greater risk of complications at similar levels of hyperglycemia compared with patients with diabetes. Objective: To describe the association between perioperative hyperglycemia and surgical complications in a population of surgical patients without diabetes receiving routine blood glucose testing and insulin administration and to evaluate the potential correlation of perioperative hyperglycemia. Design, Setting, and Participants: This retrospective cohort study of National Surgical Quality Improvement Program-defined complications after operation took place at a single academic medical center hospital from January 2013 to October 2016. Consecutive patients undergoing general, vascular, and gynecologic operations who were expected to have at least a 48-hour admission were included. Hyperglycemia was defined as blood glucose level of 140 mg/dL or higher within 24 hours after surgery. Multivariate regression was used to assess the association of hyperglycemia and complications, stratified by hyperglycemia severity and adjusted for diabetes status. Analysis began in February 2022. Exposures: Routine blood glucose testing and insulin administration. Main Outcomes and Measures: The main outcomes are odds of experiencing perioperative hyperglycemia and postoperative complication, comparing patients with and without diabetes. Results: A total of 7634 patients (mean [SD] age, 53.5 [15.1] years; 6664 patients without diabetes [83.3%] and 970 patients with diabetes [17.7%]) underwent general (6204 [81.3%]), vascular (208 [2.7%]), and gynecologic (1222 [16%]) operations. Of these, 5868 (77%) had blood glucose testing (4899 individuals without diabetes [73.5%] and 969 [99.9%] with diabetes). Hyperglycemia occurred in 882 patients with diabetes (91%) and 2484 patients without diabetes (50.7%). Of those with blood glucose level more than 180 mg/dL, 1388 (72.7%) received insulin (658 patients with diabetes who had hyperglycemia [91%] and 680 patients without diabetes who had hyperglycemia [61%]). Adjusted odds of experiencing a complication were 83% greater for patients without vs with diabetes at blood glucose level of 140 to 179 mg/dL (odds ratio, 1.83 [95% CI, 0.93-3.6]), 49% greater for blood glucose level of 180 to 249 mg/dL (odds ratio, 1.49 [95% CI, 1.06-2.11]), and 88% greater for blood glucose level more than 250 mg/dL (odds ratio, 1.88 [95% CI, 1.11-3.17]). A similar trend was observed for serious complications. Insulin may mitigate the association of hyperglycemia and complications in patients without diabetes. Conclusions and Relevance: In this study, with near universal blood glucose testing and frequent insulin use, patients without diabetes paradoxically had worse outcomes than patients with diabetes at similar levels of hyperglycemia. Insulin may mitigate this effect and broader use may improve outcomes.

Salvage regional therapy using hepatic artery infusion pump in unresectable chemotherapy resistant colorectal liver metastases.

BACKGROUND: Little is known about the influence of hepatic artery infusion pump (HAIP) therapy in the setting of chemotherapy resistant hepatic disease in the era of modern systemic therapies. METHODS: Patients who underwent HAIP therapy for chemotherapy resistant and unresectable colorectal liver metastases (CRLM) were reviewed retrospectively. RESULTS: A total of 25 patients met inclusion criteria. 52% had isolated CRLM and 92% had five or more metastatic lesions. Partial response was noted in 40% of patients. Median hepatic progression-free survival (PFS) was 7 months in those with extrahepatic disease versus 6 months in those with isolated CRLM at the time of HAIP placement (p = 0.75). Median overall survival was 8 months in patients with extrahepatic disease and 14 months in patients with isolated CRLM (p = 0.06). CONCLUSIONS: Our findings are comparable to published data and augment the literature which supports HAIP use in chemotherapy-resistant, liver-predominant metastatic colorectal cancer patients.

Patterns of Whole Exome Sequencing in Resected Cholangiocarcinoma.

BACKGROUND: With minimally effective chemotherapy options, cholangiocarcinoma patients have 5 year survival rate of 10%. Tumor genetic profiling (TGP) can identify mutations susceptible to targeted therapies. We sought to describe the use of TGP and frequency of actionable results in resected cholangiocarcinoma. METHODS: A retrospective review of patients undergoing curative intent resection at a comprehensive cancer center (2010-2020). Clinicopathologic and partial or whole exome sequencing data were reviewed. RESULTS: 114 patients (mean age 65 ± 11 years, 45% female) underwent resection of cholangiocarcinoma (46% poorly differentiated, 54% intrahepatic, 36% node positive, 75% margin negative). Additionally, 32% of patients underwent TGP, yielding a mean of 3.1 actionable mutations per patient (range 0-14). Mutations aligned with a median of one drug per patient (range 0-11). Common mutations included TP53 (33%), KRAS (31%), IDH1/2 (14%), FGFR (14%), and BRAF (8%). Targeted therapies were administered in only 4% of patients (23% of eligible sequenced patients). After a median 22 months, 23% had recurrence and 29% were deceased. DISCUSSION: TGP for cholangiocarcinoma has increased over the last decade with targeted therapies identified in most sequenced tumors, impacting treatment in a quarter of eligible patients. Precision medicine will play a central role in the future care of cholangiocarcinoma.

Colorectal cancer ovarian metastases.

OBJECTIVE: Ovarian metastases occur in 3%-5% of patients with colorectal cancer. The role of oophorectomy in that setting continues to be debated. We aimed to assess the survival of women treated with metastasectomy for ovarian metastasis. METHODS: Retrospective cohort study of patients in the California Cancer Registry (2000-2012) with stage IV colorectal cancer and ovarian metastases. Pathology other than adenocarcinoma was excluded. Adjusted Cox-proportional hazard analysis was applied to assess the risk of death. RESULTS: A total of 756 patients with synchronous ovarian metastases and 516 patients with metachronous ovarian metastases form the basis of this analysis. Median follow-up for the synchronous cohort was 21 months (IQR: 8-36). Median overall survival was 23 months (IQR: 10-42). Estimated 5-year survival reached 17% and 10-year survival was 8%. There was a significant difference in unadjusted survival between patients with solitary ovarian metastasis (median overall survival: 51 months) compared with those who had both ovarian and extraovarian metastases (20 months) (log-rank test, P<0.0001). For patients with solitary ovarian metastases, the 5- and 10-year survival was 46% and 31%, respectively. Among patients with synchronous ovarian metastases, longer unadjusted survival was observed after oophorectomy (median overall survival: 24 months) compared with no oophorectomy (18 months, log-rank P=0.01). For patients with metachronous diagnoses of colorectal cancer ovarian metastasis, the median disease-free survival was 19 months. The median survival after resection of metachronous ovarian metastases was 25 months, with the survival directly related to the disease-free interval until metastasis. For patients with resected metachronous ovarian metastases, the 5- and 10-year post-metastasectomy survival was 14% and 5%, respectively. CONCLUSIONS: Patients with colorectal cancer ovarian metastasis have favorable long-term survival. Survival rates are higher if the tumor is isolated to the ovary or if metachronous to the primary cancer.

Simulation-Based Training in Robotic Surgery: Contemporary and Future Methods.

While robotic surgery has grown in popularity and scope over the past decade, there is a persistent need for simulation-based training as surgeons adapt from the working at the bedside to the immersive and multisensory tasks at the console. From dry laboratory to virtual reality (VR) environments, simulation can be used to train surgeons in basic tasks, complex operative steps, and coordination of whole operations with members of the entire operating room (OR) staff. By integrating simulation into mentored training programs, surgeons can reduce the number of cases required to master a complex operation. Future VR based simulation will become essential to the adaptation of the surgical workforce to new technologies and adoption of emerging robotic platforms. Ultimately, robotic simulation will set standards for credentialing of new surgeons.

Does surgery provide a survival advantage in non-disseminated poorly differentiated gastroenteropancreatic neuroendocrine neoplasms?

BACKGROUND: Attributable to the high likelihood of developing distant metastatic disease, resection of poorly differentiated gastroenteropancreatic neuroendocrine neoplasms is generally contraindicated. Some patients with no distant metastatic disease will nonetheless undergo surgical resection and their outcomes are not known. We aimed to determine whether surgery confers survival advantage over systemic therapy alone for patients with non-metastatic poorly differentiated gastroenteropancreatic neuroendocrine neoplasms. METHODS: We performed a retrospective cohort study (2000-2012) of adults in the California Cancer Registry who had poorly differentiated gastroenteropancreatic neuroendocrine neoplasms (World Health Organization Grade 3) and no clinical evidence of distant metastasis (M0). Patients who underwent surgery were compared with those managed non-operatively. The adjusted Cox proportional hazards model was used to assess the risk of death. RESULTS: Among 2,245 patients (45% female, 21% pancreatic, 79% gastrointestinal), 1,549 (69%) were treated with surgery, and 696 (31%) received either systemic therapy or palliative measures alone. Median survival was 31 months after surgery versus 9 months after non-operative therapy (log-rank test, P < .001). Rates of 5-year overall survival were 39% after surgery versus 10% in the non-operative group. Adjusting for age, sex, comorbidities, receipt of chemotherapy, and tumor size and location, patients treated with surgery had a 58% lower likelihood of death compared with non-operative therapy (hazard ratio: 0.42, 95% confidence interval: 0.36-0.50, P < .001). Restricting our results to those patients who were found to have no distant metastasis intraoperatively (ie, pathologically M0), 5-year survival after surgery reached 44%. CONCLUSION: While poorly differentiated gastroenteropancreatic neuroendocrine neoplasms carries a poor prognosis, for patients with no evidence of metastatic disease, resection appears to confer significant improvement in long-term survival. Although caution and an individualized approach in treating poorly differentiated gastroenteropancreatic neuroendocrine neoplasms is advised, future guidelines might reflect this survival advantage.

Elective Surgery for Diverticulitis and the Risk of Recurrence and Ostomy.

OBJECTIVE: To assess the comparative risk of recurrence and ostomy after elective resection or medical therapy for uncomplicated diverticulitis, incorporating outpatient episodes of recurrence. BACKGROUND: While surgeons historically recommended colon resection for uncomplicated diverticulitis to reduce the risk of recurrence or colostomy, no prior studies have quantified this risk when considering outpatient episodes of disease. It remains to be determined whether surgery actually decreases those risks. METHODS: Retrospective cohort study employing an adjusted time-to-event analysis to assess the relationship of medical or surgical treatment with diverticulitis recurrence and/or receipt of an ostomy. Subjects were adults with ≥1 year continuous enrollment treated for ≥2 episodes of uncomplicated diverticulitis from a nationwide commercial claims dataset (2008-2014). RESULTS: Of 12,073 patients (mean age 56 ±â€Š14 yr, 59% women), 19% underwent elective surgery and 81% were treated by medical therapy on their second treatment encounter for uncomplicated diverticulitis. At 1 year, patients treated by elective surgery had lower rates of recurrence (6%) versus those treated by medical therapy (32%) [15% vs 61% at 5 years, adjusted hazard ratio 0.17 (95% confidence interval: 0.15-0.20)]. At 1 year, the rate of ostomy after both treatments was low [surgery (inclusive of stoma related to the elective colectomy), 4.0%; medical therapy, 1.6%]. CONCLUSIONS: Elective resection for uncomplicated diverticulitis decreases the risk of recurrence, still 6% to 15% will recur within 5 years of surgery. The risk of ostomy is not lower after elective resection, and considering colostomies related to resection, ostomy prevention should not be considered an appropriate indication for elective surgery.

Management Considerations for the Surgical Treatment of Colorectal Cancer During the Global Covid-19 Pandemic.

OBJECTIVE: The COVID-19 pandemic requires to conscientiously weigh "timely surgical intervention" for colorectal cancer against efforts to conserve hospital resources and protect patients and health care providers. SUMMARY BACKGROUND DATA: Professional societies provided ad-hoc guidance at the outset of the COVID-19 pandemic on deferral of surgical and perioperative interventions, but these lack specific parameters to determine the optimal timing of surgery. METHODS: Using the GRADE system, published evidence was analyzed to generate weighted statements for stage, site, acuity of presentation, and hospital setting to specify when surgery should be pursued, the time and duration of oncologically acceptable delays, and when to utilize nonsurgical modalities to bridge the waiting period. RESULTS: Colorectal cancer surgeries-prioritized as emergency, urgent with imminent emergency or oncologically urgent, or elective-were matched against the phases of the pandemic. Surgery in COVID-19-positive patients must be avoided. Emergent and imminent emergent cases should mostly proceed unless resources are exhausted. Standard practices allow for postponement of elective cases and deferral to nonsurgical modalities of stage II/III rectal and metastatic colorectal cancer. Oncologically urgent cases may be delayed for 6(-12) weeks without jeopardizing oncological outcomes. Outside established principles, administration of nonsurgical modalities is not justified and increases the vulnerability of patients. CONCLUSIONS: The COVID-19 pandemic has stressed already limited health care resources and forced rationing, triage, and prioritization of care in general, specifically of surgical interventions. Established guidelines allow for modifications of optimal timing and type of surgery for colorectal cancer during an unrelated pandemic.

Personal Protective Equipment and COVID-19: A Review for Surgeons.

: There is a long history of personal protective equipment (PPE) used by the surgeon to minimize the transmission of various pathogens. In the context of the present coronavirus disease 2019 pandemic there is significant controversy as to what forms of PPE are appropriate or adequate. This review aims to describe the pathogenic mechanism and route of spread of the causative virus, severe acute respiratory syndrome coronavirus, as it pertains to accumulated published data from experienced centers globally. The various forms of PPE that are both available and appropriate are addressed. There are options in the form of eyewear, gloves, masks, respirators, and gowns. The logical and practical utilization of these should be data driven and evolve based on both experience and data. Last, situations specific to surgical populations are addressed. We aim to provide granular collective data that has thus far been published and that can be used as a reference for optimal PPE choices in the perioperative setting for surgical teams.

The Safety of Expectant Management for Adhesive Small Bowel Obstruction: A Systematic Review.

BACKGROUND: Surgical training has long been to "never let the sun set on a bowel obstruction" without an operation to rule out and/or treat compromised bowel. However, advances in diagnostics have called into question the appropriate timing of non-emergent operations and expectant management is increasingly used. We performed a systematic review to evaluate the safety and effectiveness of expectant management for adhesive small bowel obstruction (aSBO) compared to early, non-emergent operation. MATERIALS & METHODS: We queried PubMed, EMBASE, and Cochrane databases for studies (1990-present) comparing early, non-emergent operations and expectant management for aSBO (PROSPERO #CRD42017057676). RESULTS: Of 4873 studies, 29 cohort studies were included for full-text review. Four studies directly compared early surgery with expectant management, but none excluded patients who underwent emergent operations from those having early non-emergent surgery, precluding a direct comparison of the two treatment types of interest. When aggregated, the rate of bowel resection was 29% in patients undergoing early operation vs. 10% in those undergoing expectant management. The rate of successful, non-operative management in the expectant group was 58%. There was a 1.3-day difference in LOS favoring expectant management (LOS 9.7 vs. 8.4 days), and the rate of death was 2% in both groups. CONCLUSION: Despite the shift towards expectant management of aSBO, no published studies have yet compared early, non-emergent operation and expectant management. A major limitation in evaluating the outcomes of these approaches using existing studies is confounding by indication related to including patients with emergent indications for surgery on admission in the early operative group. A future study, randomizing patients to early non-emergent surgery or expectant management, should inform the comparative safety and value of these approaches.

Standardization of perioperative care facilitates safe discharge by postoperative day five after pancreaticoduodenectomy.

INTRODUCTION: Pancreaticoduodenectomy is a complex surgical procedure associated with high morbidity and prolonged length of stay. Enhanced recovery after surgery principles have reduced complications rate and length of stay for multiple types of operations. We hypothesized that implementation of a standardized perioperative care pathway would facilitate safe discharge by five days after pancreaticoduodenectomy. METHODS: We performed a retrospective cohort study of patients undergoing pancreaticoduodenectomy 18 months prior to and 18 months following implementation of a perioperative care pathway at a quaternary center performing high volume pancreatic surgery. RESULTS: A total of 145 patients underwent pancreaticoduodenectomy (mean age 63 ± 10 years, 52% female), 81 before and 64 following pathway implementation, and the groups were similar in terms of preoperative comorbidities. The percentage of patients discharged within 5 days of surgery increased from 36% to 64% following pathway implementation (p = 0.001), with no observed differences in post-operative serious adverse events (p = 0.34), pancreatic fistula grade B or C (p = 0.28 and p = 0.27 respectively), or delayed gastric emptying (p = 0.46). Multivariate regression analysis showed length of stay ≤5 days three times more likely after pathway implementation. Rates of readmission within 30 days (20% pre- vs. 22% post-pathway (p = 0.75)) and 90 days (27% pre- vs. 36% post-pathway (p = 0.27)) were unchanged after pathway implementation, and were no different between patients discharged before or after day 5 at both 30 days (19% ≤5 days vs. 23% ≥ 6 days (p = 0.68)) and 90 days (32% ≤5 days vs. 30% ≥ 6 days (p = 0.81)). CONCLUSIONS: Standardizing perioperative care via enhanced recovery protocols for patients undergoing pancreaticoduodenectomy facilitates safe discharge by post-operative day five.

Challenges in Predicting Recurrence After Resection of Node-Negative Non-Small Cell Lung Cancer.

BACKGROUND: One in 5 patients with completely resected early-stage non-small cell lung cancer will recur within 2 years. Risk stratification may facilitate a personalized approach to the use of adjuvant therapy and surveillance imaging. We developed a prediction model for recurrence based on five clinical variables (tumor size and grade, visceral pleural and lymphovascular invasion, and sublobar resection), and tested the hypothesis that the addition of several new molecular markers of poor long-term outcome (vascular endothelial growth factor C; microRNA precursors 486 and 30d) would enhance prediction. METHODS: We performed a retrospective cohort study of patients with completely resected, node-negative non-small cell lung cancer from 2011 to 2014 (follow-up through 2016) using the Lung Cancer Biospecimen Resource Network. Cox regression was used to estimate the 2-year risk of recurrence. Our primary measure of model performance was the optimism-corrected C statistic. RESULTS: Among 173 patients (mean tumor size, 3.6 cm; 12% sublobar resection, 32% poorly differentiated, 16% lymphovascular invasion, 26% visceral pleural invasion), the 2-year recurrence rate was 23% (95% confidence interval, 17% to 31%). A prediction model using five known risk factors for recurrence performed only slightly better than chance in predicting recurrence (optimism-corrected C statistic, 0.54; 95% confidence interval, 0.51 to 0.68). The addition of biomarkers did not improve the model's ability to predict recurrence (corrected C statistic, 0.55; 95% confidence interval, 0.52 to 0.71). CONCLUSIONS: We were unable to predict lung cancer recurrence using a risk-prediction model based on five well-known clinical risk factors and several biomarkers. Further research should consider novel predictors of recurrence to stratify patients with completely resected early-stage non-small cell lung cancer according to their risk of recurrence.

Predicting Future Elective Colon Resection for Diverticulitis Using Patterns of Health Care Utilization.

BACKGROUND: Recurrent diverticulitis is the most common reason for elective colon surgery and, although professional societies now recommend against early resection, its use continues to rise. Shared decision making decreases use of low-value surgery but identifying which patients are most likely to elect surgery has proven difficult. We hypothesized that Machine Learning algorithms using health care utilization (HCU) data can predict future clinical events including early resection for diverticulitis. STUDY DESIGN: We developed models for predicting future surgery among patients with new diagnoses of diverticulitis (2009-2012) from the MarketScan® database. Claims data (diagnosis, procedural, and drug codes) were used to train three Machine Learning algorithms to predict surgery occurring between 52 and 104 weeks following diagnosis. RESULTS: Of 82,231 patients with incident diverticulitis (age 51 ± 8 years, 52% female), 1.2% went on to elective colon resection. Using maximal training data (152 consecutive weeks of claims), the Gradient Boosting Machine model predicted elective surgery with an area under the curve (AUC) of 75% (95% uncertainty interval [UI] 71-79%). Models trained on less data resulted in less accurate prediction (AUC: 68% [64-74%] using 128 weeks, 57% [53-63%] using 104 weeks). The majority of resections (85%) were identified as low-value. CONCLUSION: By applying Machine Learning to HCU data from the time around a diagnosis of diverticulitis, we predicted elective surgery weeks to months in advance, with moderate accuracy. Identifying patients who are most likely to elect surgery for diverticulitis provides an opportunity for effective shared decision making initiatives aimed at reducing the use of costly low-value care.

Variability in invasive mediastinal staging for lung cancer: A multicenter regional study.

OBJECTIVE: Prior studies have reported underuse of-but not variability in-invasive mediastinal staging in the pretreatment evaluation of patients with lung cancer. We sought to compare rates of invasive mediastinal staging for lung cancer across hospitals participating in a regional quality improvement and research collaborative. METHODS: We conducted a retrospective study (2011-2013) of patients undergoing resected lung cancer from the Surgical Clinical Outcomes and Assessment Program in Washington State. Invasive mediastinal staging included mediastinoscopy and/or endobronchial/esophageal ultrasound-guided nodal aspiration. We used a mixed-effects model to mitigate the influence of small sample sizes at any 1 hospital on rates of invasive staging and to adjust for hospital-level differences in the frequency of clinical stage IA disease. RESULTS: A total of 406 patients (mean age, 68 years; 69% clinical stage IA; and 67% lobectomy) underwent resection at 5 hospitals (4 community and 1 academic). Invasive staging occurred in 66% of patients (95% confidence interval [CI], 61%-71%). CI inspection revealed that 2 hospitals performed invasive staging significantly more often than the overall average (94%, [95% CI, 89%-96%] and 84% [95% CI, 78%-88%]), whereas 2 hospitals performed invasive staging significantly less often than overall average (31% [95% CI, 21%-44%] and 17% [95% CI, 7%-36%]). CONCLUSIONS: Rates of invasive mediastinal staging varied significantly across hospitals providing surgical care for patients with lung cancer. Future studies that aim to understand the reasons underlying variability in care may inform quality improvement initiatives or lead to the development of novel staging algorithms.