Accurate Determination of Colorectal Polyp Counts Is Not Mission Impossible for Pathologists and Is Essential for Surveillance Interval Decision-Making: A Review of 1294 Specimens

Introduction/Objective

The US Multi-Society Task Force on colorectal cancer updated recommendations in 2020 for follow-up after colonoscopy. This emphasized differential surveillance intervals based on the number of polyps removed. We determined how often we accurately counted the number of polyps and what submission practices allowed accurate quantification.

Methods/Case Report

We reviewed consecutive colorectal polyp cases from 10/1 to 12/31, 2019, excluding inflammatory polyps in inflammatory bowel disease. Pathology reports were reviewed to determine if the polyp number could be determined from history, gross description, and histologic findings. When numbers did not match or were unclear, endoscopy reports and glass slides were reviewed.

Results (if a Case Study enter NA)

601 cases with 1294 specimens were identified. We accurately determined polyp number in 1235 (95.4%) specimens. 1072 (82.8%) specimens were submitted as single polyp per container (SP), while 222 (17.2%) were submitted with > 1 polyp per container (MP). Polyp number was not quantifiable in 58 (26.1%) MP due to 11 labeled as “multiple polyps” in requisition/endoscopy reports and 47 submitted with polyp number different from number of tissue portions seen grossly and microscopically. In 1 (0.1%) SP, polyp number was unclear because separate portions of tissue contained two different diagnoses. Rate of accurate polyp count is significantly different between SP and MP (p<0.01) by Fisher exact test.

Conclusion

We accurately determined number of polyps in most specimens (95.4%) because our gastroenterologists usually submit SP. In MP, polyp count in 26.1% of specimens was unquantifiable. This could lead to uncertainty in surveillance follow-up intervals. Therefore, we recommend submitting one polyp per container.

Utility of machine learning in developing a predictive model for early-age-onset colorectal neoplasia using electronic health records

BACKGROUND AND AIMS

The incidence of colorectal cancer (CRC) is increasing in adults younger than 50, and early screening remains challenging due to cost and under-utilization. To identify individuals aged 35-50 years who may benefit from early screening, we developed a prediction model using machine learning and electronic health record (EHR)-derived factors.

METHODS

We enrolled 3,116 adults aged 35-50 at average-risk for CRC and underwent colonoscopy between 2017-2020 at a single center. Prediction outcomes were (1) CRC and (2) CRC or high-risk polyps. We derived our predictors from EHRs (e.g., demographics, obesity, laboratory values, medications, and zip code-derived factors). We constructed four machine learning-based models using a training set (random sample of 70% of participants): regularized discriminant analysis, random forest, neural network, and gradient boosting decision tree. In the testing set (remaining 30% of participants), we measured predictive performance by comparing C-statistics to a reference model (logistic regression).

RESULTS

The study sample was 55.1% female, 32.8% non-white, and included 16 (0.05%) CRC cases and 478 (15.3%) cases of CRC or high-risk polyps. All machine learning models predicted CRC with higher discriminative ability compared to the reference model [e.g., C-statistics (95%CI); neural network: 0.75 (0.48-1.00) vs. reference: 0.43 (0.18-0.67); P = 0.07] Furthermore, all machine learning approaches, except for gradient boosting, predicted CRC or high-risk polyps significantly better than the reference model [e.g., C-statistics (95%CI); regularized discriminant analysis: 0.64 (0.59-0.69) vs. reference: 0.55 (0.50-0.59); P<0.0015]. The most important predictive variables in the regularized discriminant analysis model for CRC or high-risk polyps were income per zip code, the colonoscopy indication, and body mass index quartiles.

DISCUSSION

Machine learning can predict CRC risk in adults aged 35-50 using EHR with improved discrimination. Further development of our model is needed, followed by validation in a primary-care setting, before clinical application.

Abstract CT236: A phase 1b, multicenter, randomized, blinded, placebo-controlled study to evaluate the efficacy of guselkumab in subjects with familial adenomatous polyposis

Background: Familial adenomatous polyposis (FAP) is the most common hereditary polyposis syndrome. It is an autosomal dominant inherited disorder characterized by the early onset of hundreds to thousands of adenomatous polyps throughout the colon. If left untreated, nearly all individuals with this syndrome develop colorectal cancer (CRC) by the third decade of life. Prophylactic colectomy is the standard of care, but individuals remain at risk for malignant transformation of duodenal polyps, rectal polyps for those who have undergone rectal-sparing surgeries, and ileal pouch polyps for those with ileal pouch-anal anastomoses. Multiple studies with both nonselective and selective cyclooxygenase inhibitors (such as sulindac or celecoxib) have shown that anti-inflammatory agents may prevent the formation and inhibit the growth of colorectal adenomatous polyps. However, toxicities associated with these agents and their limited efficacy have prevented their further development. Therefore, there is a high unmet need for novel treatment options to reduce polyp burden, delay or eliminate the need for colectomy and recurrent rectal surgery, and intercept the development of adenocarcinomas in individuals with FAP. Polyps from individuals with FAP display inflammatory features associated with the activation of the IL-23/IL-17/JAK/STAT3 pathway. This inflammation is thought to contribute to further carcinogenesis, culminating in tumor development. Specifically, IL-23 is linked to tumor growth and progression in CRC, and adenomas with high-grade dysplasia showed elevated levels of IL-17A and pSTAT3. Guselkumab, a human monoclonal antibody directed against the p19 subunit of IL-23, specifically targets IL-23 and inhibits its interaction with the IL-23 receptor. Pre-clinical models suggest that inhibition of IL-23 signaling will result in less inflammation and reduce tumor development. Methods: This randomized, blind, placebo-controlled study will evaluate the safety and efficacy of guselkumab in adults with FAP (genetic or clinical diagnosis) who have already undergone colectomy. Polyps with a sum of diameters ≥10 mm in the rectum or pouch are required. Subjects will be randomized equally to one of three study arms: 100 mg, 300 mg, or placebo given subcutaneously every 4 weeks for 6 doses. The primary efficacy endpoint is percentage change from baseline in rectal/pouch polyp burden after 24 weeks. Secondary efficacy endpoints include duodenal polyp burden change and changes in InSiGHT and Spigelman staging. Exploratory translational research objectives will explore changes in RNA expression profiles, epigenomic profiles, cytokine levels, and the microbiome. Exclusion criteria include prior IL-23 targeted therapies and any polyps &gt;1 cm that cannot be removed. While participating, subjects are required to stop any other FAP-directed drug therapy except for aspirin. As of January 2019, 18% of the planned 72 subjects have been enrolled. ClinicalTrials.gov Identifier: NCT03649971.

Citation Format: Eduardo Vilar-Sanchez, Carol Burke, Marcia R. Cruz Correa, Evelien Dekker, William M. Grady, Philippe Grandval, Bryson W. Katona, Xavier Llor, Douglas L. Riegert-Johnson, Jean-Christophe Saurin, Peter Stanich, Daniel A. Sussman, David Weinberg, Edward F. Attiyeh, Devanand Joseph, Kelly Raybold, Gary V. Borzillo, Thomas J. Prior, Michael Smith, Hong Xie, Kurtis E. Bachman, Jeffrey R. Infante, Niloy Jewel Samadder. A phase 1b, multicenter, randomized, blinded, placebo-controlled study to evaluate the efficacy of guselkumab in subjects with familial adenomatous polyposis [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr CT236.

A phase II study of PD-1 inhibition for the prevention of colon adenomas in patients with Lynch syndrome and a history of partial colectomy

TPS1587

Background: Colon cancers and adenomas that are associated with Lynch syndrome (LS) often display microsatellite instability (MSI), a characteristic that is associated with increased response to treatment with PD-1 inhibitors. Because LS patients with a history of colon cancer are at increased risk of having a second primary colon cancer or high-risk adenoma, preventive measures are of particular interest in this population. We hypothesize that a maintenance schedule of nivolumab can be safely administered to LS patients with a history of treated colon cancer with remaining colon at risk in order to decrease the incidence of adenomas, advanced adenomas and second primary colon cancers. Methods: OSU 17198 is a phase II multi-center, single-arm study of nivolumab in patients with germline MLH1 or MSH2 mutations and a history of hemicolectomy for colon cancer at least one year prior to study entry. Subjects must have completed any adjuvant therapy at least 6 months prior to study participation and may not have received prior therapy with a PD-1 inhibitor. Nivolumab is given at 240mg IV every 3 months for two years, and colonoscopies will be performed prior to study entry, after the fourth dose, after the eighth dose, and one year after the eighth dose. Subjects will be monitored for auto-immune adverse effects. The primary endpoint is incidence of adenomas at three years, and secondary endpoints are safety, incidence of advanced adenomas, and incidence of colon and non-colon cancers at three years. Approximately 104 subjects will be enrolled to obtain 94 evaluable subjects. This study is currently open for enrollment at the Ohio State University and at various stages of activation at seven additional sites in the United States. Enrollment of this study is anticipated to be completed in 2020, and data collection is anticipated to be complete in 2023. This study has undergone safety review by the FDA and the Ohio State University Institutional Review Board. Clinical trial information: NCT03631641.

Improvement of children's nutritional status after enteral feeding by PEG: an interim report

BACKGROUND

Enteral feeding by percutaneous endoscopic gastrostomy is indicated as long-term nutritional support for children with neurologic impairment and dysphagia. We report our experience with percutaneous endoscopic gastrostomy and evaluate the age range of children with cerebral palsy who benefit most with weight and height gain.

METHODS

In a prospective study, from August 1996 to August 1997, 20 endoscopic gastrostomies were performed in 20 children diagnosed with cerebral palsy (16), myopathy (2), and brain trauma (2). The mean age was 6.5 years and the mean follow-up 5.9 months. All patients had severe mental impairment and oropharyngeal dysphagia. They were followed up monthly on an outpatient basis by both the gastroenterologist and the dietitian, who assessed gastrostomy complications and performed anthropometric measurements.

RESULTS

All 20 patients benefited from enteral nutrition with a statistically significant gain in weight (p < 0.01), and there was a trend toward improved weight/height ratio in children under 4 years of age according to Z-score and mid-arm muscle area (p < 0.01). Triceps skinfold thickness failed to reach statistical significance. There were no immediate complications related to the procedure. Perforations occurred with three (15%) tubes, and the plugs for introduction of food had to be replaced after 4 months of use. All complications, namely formation of granulation tissue at the stoma (7), stoma infection (4), gastroesophageal reflux pneumonia (3), and pneumoperitoneum (1) were managed clinically.

CONCLUSIONS

Endoscopic gastrostomy is a safe procedure for children. Enteral feeding resulted in a trend toward a normalized weight/height ratio for children with cerebral palsy younger than 4 years and significant weight gain in those older than 12 years.

Nutritional status improvement in neurologically impaired patients by percutaneous endoscopic gastrostomy feeding

With increased use of percutaneous endoscopic gastrostomy, it became clear that neurologically impaired patients might benefit from its use. From August 1996 to July 1997, we performed 19 percutaneous endoscopic gastrostomies in patients with neurological sequelae, who were incapable to maintain their nutritional status by oral ingestion or had repeated episodes of aspiration. Sixteen patients were followed prospectively, from 30 days to 11 months (median: 6.4 months). Average weight (38.2 kg to 44.8 kg), BMI (14.8 kg/m2 to 17.8 kg/m2), weight/height ratio (23.5 kg/cm to 28 kg/cm), mid-upper arm circumference (19.4 cm to 21.6 cm) and triceps skinfold thickness (10.3 mm to 12.6 mm) were significantly increased (P < 0.01). Before percutaneous endoscopic gastrostomy, there were 10 (10/16, 62.5%) patients with grade III thinness. In this group, 3/10 patients (30%) showed improvement to grade I (two individuals) and II (one patient). All but five patients (68.75%) were below the fifth percentile of normal distribution for mid-upper arm circumference. One patient (6.2%) showed improvement of her status (between 25th and 50th percentiles). Four patients (25%) started the follow-up below the fifth percentile for normal distribution of triceps skinfold thickness, and showed no improvement. There were no early complications secondary the procedure. Late complications included granulation tissue on ostomy site (18.8%) and ostomy infection (6.2%). Statistical analysis showed significant improvement of anthropometric data. Percutaneous endoscopic gastrostomy is a simple, highly successful and safe procedure, when performed in neurologically impaired patients. It is efficient as a long-term enteral feeding method.