Early Postoperative Prediction of Complications and Readmission After Colorectal Cancer Surgery Using an Artificial Neural Network.

BACKGROUND: Early predictors of postoperative complications can risk-stratify patients undergoing colorectal cancer surgery. However, conventional regression models have limited power to identify complex nonlinear relationships among a large set of variables. We developed artificial neural network models to optimize the prediction of major postoperative complications and risk of readmission in patients undergoing colorectal cancer surgery. OBJECTIVE: The aim of this study was to develop an artificial neural network model to predict postoperative complications using postoperative laboratory values, and compare these models' accuracy to standard regression methods. DESIGN: This retrospective study included patients who underwent elective colorectal cancer resection between January 1, 2016, and July 31, 2021. Clinical data, cancer stage, and laboratory data from postoperative day 1 to 3 were collected. Models of complications and readmission risk were created using multivariable logistic regression and single-layer neural networks. SETTING: National Cancer Institute-Designated Comprehensive Cancer Center. PATIENTS: Adult colorectal cancer patients. MAIN OUTCOME MEASURES: Accuracy of predicting postoperative major complication, readmission and anastomotic leak using the area under the receiver-operating characteristic curve. RESULTS: Neural networks had larger areas under the curve for predicting major complications compared to regression models (neural network 0.811; regression model 0.724, p < 0.001). Neural networks also showed an advantage in predicting anastomotic leak (p = 0.036) and readmission using postoperative day 1-2 values (p = 0.014). LIMITATIONS: Single-center, retrospective design limited to cancer operations. CONCLUSIONS: In this study, we generated a set of models for early prediction of complications after colorectal surgery. The neural network models provided greater discrimination than the models based on traditional logistic regression. These models may allow for early detection of postoperative complications as soon as postoperative day 2. See Video Abstract.

Endoscopic and imaging outcomes of PD-1 therapy in localised dMMR colorectal cancer.

BACKGROUND: Neoadjuvant immune checkpoint blockade (IO) is emerging as a therapeutic option for patients with deficient mismatch repair (dMMR) colorectal cancer (CRC) given high pathological response rates. The aim of the study was to characterise imaging and endoscopic response to IO. METHODS: A retrospective analysis of patients with localised dMMR CRC that received at least one cycle of neoadjuvant anti-PD-1 therapy was conducted. Endoscopy, imaging, and pathological outcomes were reviewed to determine response to treatment according to standardised criteria. RESULTS: Thirty-eight patients had received IO for the treatment of localised CRC (median eight cycles). Among evaluable cases (n = 31 for endoscopy and n = 34 for imaging), the best endoscopic response was complete response (CR) in 45% of cases, and the best radiographic response was CR in 23% of cases. Imaging CR rate after ≤4 cycles of IO (n = 1) was 6% compared to 44% after >4 IO cycles (n = 7). Among 28 patients with imaging and endoscopy available, a discrepancy in best response was noted in 15 (54%) cases. At a median follow-up of 28.2 months from IO start, 18 patients underwent surgical resection of which 11 (61%) had pathological CR (pCR). Despite pCR or no evidence of progression ≥6 months after completion of IO among non-operatively managed patients, 72% and 42% of patients had non-CR on imaging and endoscopy, respectively. CONCLUSIONS: Discrepancies between imaging and endoscopy are prevalent, and irregularities identified on these modalities can be identified despite pathological remission. Improved clinical response criteria are warranted.

Optimization of an mHealth lifestyle intervention for families with hereditary cancer syndromes: Study protocol for a multiphase optimization strategy feasibility study.

BACKGROUND: Individuals at increased hereditary risk of cancer are an important target for health promotion and cancer prevention interventions. Health-4-Families uses the Multiphase Optimization STrategy (MOST) framework and is designed to pilot digital delivery strategies for a distance-based, 16-week intervention to promote weight management, healthy diet, and increased physical activity among individuals with BRCA1/BRCA2 or DNA mismatch repair (MMR) pathogenic germline variants. This communication describes participant recruitment and the design of the Health-4-Families pilot study. METHODS: Health-4-Families is a full-factorial (16 condition) randomized pilot study of four lifestyle intervention components: social networking, telephone or email coaching, text messaging, and self-monitoring. The primary outcome was feasibility and satisfaction with these study components. Participants with pathogenic germline variants were identified via clinic surveillance lists and advocacy organizations and were invited to participate with family members. All participants had to report meeting at least one of the following criteria: (1) having a BMI ≥ 25 kg/m2, (2) consuming <5 servings of fruit and vegetables per day, or (3) getting <150 min of moderate-to-vigorous intensity activity per week. RESULTS: The majority of screened potential participants with pathogenic variants (83%) were eligible; 86% of those eligible provided informed consent and 79% (n = 104) completed baseline. A total of 206 family members were nominated by study participants and 49% (n = 102) completed baseline. DISCUSSION: Recruitment data suggest that individuals with pathogenic germline variants, who are at increased risk for hereditary cancers, are motivated to participate in digital lifestyle interventions. This recruitment success highlights the importance of identifying and prioritizing effective and efficient intervention components for hereditary cancer families. We intend to use the outcomes of our pilot study to inform a fully-powered factorial study for this community.

The Provocative Roles of Platelets in Liver Disease and Cancer.

Both platelets and the liver play important roles in the processes of coagulation and innate immunity. Platelet responses at the site of an injury are rapid; their immediate activation and structural changes minimize the loss of blood. The majority of coagulation proteins are produced by the liver-a multifunctional organ that also plays a critical role in many processes: removal of toxins and metabolism of fats, proteins, carbohydrates, and drugs. Chronic inflammation, trauma, or other causes of irreversible damage to the liver can dysregulate these pathways leading to organ and systemic abnormalities. In some cases, platelet-to-lymphocyte ratios can also be a predictor of disease outcome. An example is cirrhosis, which increases the risk of bleeding and prothrombotic events followed by activation of platelets. Along with a triggered coagulation cascade, the platelets increase the risk of pro-thrombotic events and contribute to cancer progression and metastasis. This progression and the resulting tissue destruction is physiologically comparable to a persistent, chronic wound. Various cancers, including colorectal cancer, have been associated with increased thrombocytosis, platelet activation, platelet-storage granule release, and thrombosis; anti-platelet agents can reduce cancer risk and progression. However, in cancer patients with pre-existing liver disease who are undergoing chemotherapy, the risk of thrombotic events becomes challenging to manage due to their inherent risk for bleeding. Chemotherapy, also known to induce damage to the liver, further increases the frequency of thrombotic events. Depending on individual patient risks, these factors acting together can disrupt the fragile balance between pro- and anti-coagulant processes, heightening liver thrombogenesis, and possibly providing a niche for circulating tumor cells to adhere to-thus promoting both liver metastasis and cancer-cell survival following treatment (that is, with minimal residual disease in the liver).

Long-term survival among 5-year survivors of adolescent and young adult cancer.

BACKGROUND: Although there are a growing number of survivors of adolescent and young adult (AYA) cancer, to the authors' knowledge the long-term overall survival (OS) patterns for AYA cancer survivors are underreported. The objective of the current study was to assess the long-term survival of AYA cancer survivors and identify factors associated with diminished long-term survival. METHODS: The authors used The University of Texas MD Anderson Cancer Center's tumor registry to identify 5-year survivors of cancer diagnosed as AYAs (ages 15-39 years) between the years 1970 and 2005, and who were alive 5 years after diagnosis. Kaplan-Meier curves were used to estimate OS rates over time, and Cox proportional hazards models were fitted to evaluate the association of covariates with OS. RESULTS: The authors identified 16,728 individuals who were 5-year survivors of cancer and were diagnosed as AYAs with a median follow-up of 20.0 years. The 10-year, 20-year, and 25-year OS rates were 86% (95% confidence interval [95% CI], 85%-86%), 74% (95% CI, 73%-75%), and 68% (95% CI, 67%-68%), respectively, all of which were lower than the age-adjusted estimated survival rates of the general population. Long-term OS improved for AYAs diagnosed between 2000 and 2005 compared with those diagnosed in the prior decades (P < .001). Older age at the time of diagnosis, receipt of radiation, and diagnoses including central nervous system tumors and breast cancer each were associated with diminished long-term survival. CONCLUSIONS: AYA cancer survivors have inferior long-term survival compared with the general population. Studies investigating the prevalence and types of late treatment effects and causes of death among AYA survivors are needed to more accurately identify AYAs who are at highest risk of early or late mortality.

Lynch Syndrome: A Primer for Urologists and Panel Recommendations.

PURPOSE: Lynch syndrome, also known as hereditary nonpolyposis colorectal cancer, is a common genetic disease. The predisposition of patients with Lynch syndrome to urological cancer, particularly upper tract urothelial carcinoma, is underappreciated. Urologists may be involved in several aspects of care involving Lynch syndrome, including identifying undiagnosed patients, surveillance of those with established Lynch syndrome or screening family members, in addition to treating patients with Lynch syndrome in whom upper tract urothelial carcinoma develops. We sought to increase awareness in the urological community about Lynch syndrome and provide some guidance where little currently exists. MATERIALS AND METHODS: Using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement we reviewed the available published literature and guidelines from 1998 to 2014 on Lynch syndrome and its association with upper tract urothelial carcinoma. Recommendations based on the literature and the consensus of expert opinion are provided. RESULTS: No randomized or prospective study has been done to evaluate Lynch syndrome in the setting of urological cancer. All data were based on retrospective studies. Lynch syndrome is an autosomal dominant genetic disease caused by germline mutations in 4 mismatch repair genes, leading to the accumulation of DNA errors in microsatellite regions. Upper tract urothelial carcinoma develops in up to 28% of patients with known Lynch syndrome. The diagnosis of Lynch syndrome is established by clinical criteria, tumor tissue testing and genetic evaluation. Urologists should suspect Lynch syndrome when a patient with upper tract urothelial carcinoma presents before age 60 years or meets the 3-2-1 rule. Screening patients with Lynch syndrome for upper tract urothelial carcinoma presents a particular challenge. While no ideal screening test exists, at a minimum routine urinalysis is recommended using the American Urological Association guideline of 3 or more red blood cells per high power field as a trigger for further assessment. Upper tract urothelial carcinoma associated with Lynch syndrome presents at a younger age than sporadic upper tract urothelial carcinoma. It shows a higher proportion of ureteral cancer with a female preponderance and a possible predisposition to bilaterality. CONCLUSIONS: Lynch syndrome is a common genetic disease that is an underappreciated cause of upper tract urothelial carcinoma and possibly other urological cancers. Optimal screening for upper tract urothelial carcinoma in this population is unclear. Further study is needed to identify the best screening test and interval of testing. Urologists should consider routine tissue testing of de novo upper tract urothelial carcinoma tissue in individuals at risk.

An individualized conditional survival calculator for patients with rectal cancer.

BACKGROUND: Conditional survival estimates account for time survived since diagnosis to provide prognostic information for long-term cancer survivors. For rectal cancer, stage-related treatment (eg, neoadjuvant radiotherapy) affects pathologic stage and therefore stage-associated survival estimates. OBJECTIVES: The aim of this study is to estimate conditional survival for patients who have rectal cancer and to develop an interactive calculator to use for individualized patient counseling. PATIENTS: Patients with rectal adenocarcinoma were identified by using the Surveillance Epidemiology and End Results registry (1988-2002, N = 22,610). DESIGN: Cox regression models were developed to determine adjusted survival estimates (years 1-10) and used to calculate 5-year adjusted conditional survival. Models were built separately for no radiotherapy, preoperative radiotherapy, postoperative radiotherapy, and patients with stage IV disease. Covariates included age, sex, race, tumor grade, and type of surgery. An Internet-based conditional survival calculator was developed. RESULTS: Radiotherapy was given to 42.6% of patients (14.1% preoperative, 28.4% postoperative). Significant improvements in 5-year conditional survival were observed for all stages, with the exception of stage I because of the initial high survival probability at diagnosis. Patients with advanced stage had the greatest improvements in conditional survival, with 5-year absolute increases of 33% (stage IIIC) and 54% (IV). Other factors associated with conditional survival included sequence of radiotherapy and surgery, age, race, and tumor grade. The Internet-based conditional survival calculator can be accessed at www.mdanderson.org/rectalcalculator. LIMITATIONS: The data source used does not include information on chemotherapy treatment, change in staging after neoadjuvant treatment, or patient comorbidities. CONCLUSION: Conditional survival estimates improve over 5 years in patients who have rectal cancer; the greatest improvements are observed among patients with advanced stage disease. The conditional survival calculator is an individualized decision support tool that informs patients, who must make non-treatment-related life decisions, and their clinicians planning follow-up and surveillance.