Incidence of diagnostic change in colorectal polyp specimens after deeper sectioning at 2 different laboratories staffed by the same pathologists

Accuracy of polyp characterization by artificial intelligence and endoscopists: a prospective, non-randomized study in a tertiary endoscopy center

Background and study aims Artificial intelligence (AI) in gastrointestinal endoscopy is developing very fast. Computer-aided detection of polyps and computer-aided diagnosis (CADx) for polyp characterization are available now. This study was performed to evaluate the diagnostic performance of a new commercially available CADx system in clinical practice. Patients and methods This prospective, non-randomized study was performed at a tertiary academic endoscopy center from March to August 2022. We included patients receiving a colonoscopy. Polypectomy had to be performed in all polyps. Every patient was examined concurrently by an endoscopist and AI using two opposing screens. The AI system, overseen by a second observer, was not visible to the endoscopist. The primary outcome was accuracy of the AI classifying the polyps into "neoplastic" and "non-neoplastic." The secondary outcome was accuracy of the classification by the endoscopists. Sessile serrated lesions were classified as neoplastic. Results We included 156 patients (mean age 65; 57 women) with 262 polyps ≤10 mm. Eighty-four were hyperplastic polyps (32.1%), 158 adenomas (60.3%), seven sessile serrated lesions (2.7%) and 13 other entities (normal/inflammatory colonmucosa, lymphoidic polyp) (4.9%) on histological diagnosis. Sensitivity, specificity and accuracy of AI were 89.70% (95% confidence interval [CI]: 84.02%-93.88%), 75.26% (95% CI: 65.46%-83.46%) and 84.35% (95% CI:79.38%-88.53%), respectively. Sensitivity, specificity and accuracy for less experienced endoscopists (2-5 years of endoscopy) were 95.56% (95% CI: 84.85%-99.46%), 61.54% (95% CI: 40.57%-79.77%) and 83.10% (95% CI: 72.34%-90.95%) and for experienced endoscopists 90.83% (95% CI: 84.19%-95.33%), 71.83% (95% CI: 59.90%-81.87%) and 83.77% (95% CI: 77.76%-88.70%), respectively. Conclusion Accuracy for polyp characterization by a new commercially available AI system is high, but does not fulfill the criteria for a "resect-and-discard" strategy.

Implementation of optical diagnosis with a "resect and discard" strategy in clinical practice: DISCARD3 study

BACKGROUND AND AIMS

Optical diagnosis (OD) of polyps can be performed with advanced endoscopic imaging. For high-confidence diagnoses, a "resect and discard" strategy could offer significant histopathology time and cost savings. The implementation threshold is a ≥90% OD-histology surveillance interval concordance. Here we assessed the OD learning curve and feasibility of a resect and discard strategy for ≤5-mm and <10-mm polyps in a bowel cancer screening setting.

METHODS

In this prospective feasibility study, 8 bowel cancer screening endoscopists completed a validated OD training module and performed procedures. All <10-mm consecutive polyps had white-light and narrow-band images taken and were given high- or low-confidence diagnoses until 120 high-confidence ≤5-mm polyp diagnoses had been performed. All polyps had standard histology. High-confidence OD errors underwent root-cause analysis. Histology and OD-derived surveillance intervals were calculated.

RESULTS

Of 565 invited patients, 525 patients were included. A total of 1560 <10-mm polyps underwent OD and were resected and retrieved (1329 ≤5 mm and 231 6-9 mm). There were no <10-mm polyp cancers. High-confidence OD was accurate in 81.5% of ≤5-mm and 92.8% of 6-9-mm polyps. Sensitivity for OD of a ≤5-mm adenoma was 93.0% with a positive predictive value of 90.8%. OD-histology surveillance interval concordance for ≤5-mm OD was 91.3% (209/229) for U.S. Multi-Society Task Force, 98.3% (225/229) for European Society of Gastrointestinal Endoscopy, and 98.7% (226/229) for British Society of Gastroenterology guidelines, respectively.

CONCLUSIONS

A resect and discard strategy for high-confidence ≤5-mm polyp OD in a group of bowel cancer screening colonoscopists is feasible and safe, with performance exceeding the 90% surveillance interval concordance required for implementation in clinical practice. (Clinical trial registration number: NCT04710693.).

Deeper Sections: Its Frequency and Diagnostic Utility in Histopathology of Noncutaneous Small Biopsy Specimen in a Tertiary Hospital in Nepal

BACKGROUND

Biopsy is an important tool for accurate diagnosis of disease in histopathology which can be examined at multiple levels during microscopic examination. The initial sections may not be representative of the entire biopsy, which leads to frequent request for deeper sections. This study assessed the frequency and diagnostic utility of deeper sections in noncutaneous small biopsy specimens at a tertiary hospital. Material and Methods. All the formalin fixed noncutaneous small biopsy specimens received were processed for grossing followed by tissue processing and embedding. The paraffin-embedded blocks were cut into 3-5 µm sections, fixed in a glass slide, deparaffinised, and then stained with hematoxylin and eosin (H&E) stain as per the standard protocol. Deeper sections were instructed based on microscopic findings of initial slides. The overall frequency of deeper section, its levels of cutting, tissue survival, and outcome were assessed and interpreted for the final diagnosis.

RESULTS

A total of 125 cases (26.9%) from 464 samples received were requested for deeper sections. The most frequent deeper sections were from cervix (72 cases) followed by stomach (18 cases) and endometrium (17 cases). The deeper sections were performed most frequently at 4 levels (17.8%) followed by 5 levels (14.8%) and 6 levels (13.3%). Deeper sections revealed nondiagnostic additional features in 28.2%, while 2.2% showed additional diagnostic features. Likewise, 2.2% had tissue segment loss. The overall mean level showing additional features was 6 levels. Tissue survival increased in 13% cases and 1.5% had decreased survival. The most common reason for requesting deeper section was to enhance sensitivity and diagnostic accuracy of disease.

CONCLUSIONS

Deeper sections often contribute to final diagnosis. Therefore, deeper sections on biopsies that cannot readily be diagnosed on routine levels are recommended regardless of size of the lesion and there should be uniformity in the practice of deeper sections across the globe.

Artificial intelligence in perioperative management of major gastrointestinal surgeries

Artificial intelligence (AI) demonstrated by machines is based on reinforcement learning and revolves around the usage of algorithms. The purpose of this review was to summarize concepts, the scope, applications, and limitations in major gastrointestinal surgery. This is a narrative review of the available literature on the key capabilities of AI to help anesthesiologists, surgeons, and other physicians to understand and critically evaluate ongoing and new AI applications in perioperative management. AI uses available databases called “big data” to formulate an algorithm. Analysis of other data based on these algorithms can help in early diagnosis, accurate risk assessment, intraoperative management, automated drug delivery, predicting anesthesia and surgical complications and postoperative outcomes and can thus lead to effective perioperative management as well as to reduce the cost of treatment. Perioperative physicians, anesthesiologists, and surgeons are well-positioned to help integrate AI into modern surgical practice. We all need to partner and collaborate with data scientists to collect and analyze data across all phases of perioperative care to provide clinical scenarios and context. Careful implementation and use of AI along with real-time human interpretation will revolutionize perioperative care, and is the way forward in future perioperative management of major surgery.

Hyperplastic polyp or sessile serrated lesion? The contribution of serial sections to reclassification

BACKGROUND

The histological discrimination of hyperplastic polyps from sessile serrated lesions can be difficult. Sessile serrated lesions and hyperplastic polyps are types of serrated polyps which confer different malignancy risks, and surveillance intervals, and are sometimes difficult to discriminate. Our aim was to reclassify previously diagnosed hyperplastic polyps as sessile serrated lesions or confirmed hyperplastic polyps, using additional serial sections.

METHODS

Clinicopathological data for all colorectal hyperplastic polyps diagnosed in 2016 and 2017 was collected. The slides were reviewed and classified as hyperplastic polyps, sessile serrated lesion, or other, using current World Health Organization criteria. Eight additional serial sections were performed for the confirmed hyperplastic polyp group and reviewed.

RESULTS

Of an initial 147 hyperplastic polyps from 93 patients, 9 (6.1%) were classified as sessile serrated lesions, 103 as hyperplastic polyps, and 35 as other. Of the 103 confirmed hyperplastic polyps, 7 (6.8%) were proximal, and 8 (7.8%) had a largest fragment size of ≥5 mm and < 10 mm. After 8 additional serial sections, 11 (10.7%) were reclassified as sessile serrated lesions. They were all less than 5 mm and represented 14.3% of proximal polyps and 10.4% of distal polyps. An average of 3.6 serial sections were required for a change in diagnosis.

CONCLUSION

Histopathological distinction between hyperplastic polyps and sessile serrated lesions remains a challenge. This study has uncovered a potential role for the use of additional serial sections in the morphological reappraisal of small hyperplastic polyps, especially when proximally located.

Artificial intelligence technology applications in the pathologic diagnosis of the gastrointestinal tract

Artificial intelligence (AI) is a complex technology with a steady flow of new applications, including in the pathology laboratory. Applications of AI in pathology are scarce but increasing; they are based on complex software-based machine learning with deep learning trained by pathologists. Their uses are based on tissue identification on histologic slides for classification into categories of normal, nonneoplastic and neoplastic conditions. Most AI applications are based on digital pathology. This commentary describes the role of AI in the pathological diagnosis of the gastrointestinal tract and provides insights into problems and future applications by answering four fundamental questions.

Are Upfront Levels for Colon "Polyps" Necessary? A Pragmatic Review

INTRODUCTION

Colon biopsies are among the most frequently examined specimens by pathologists. Many pathology practices, ours included, review upfront levels on all gastrointestinal biopsies. In our experience, when a lesion is present on specimens labeled "colon polyp," it is readily identified on the first level. To test our hypothesis, we re-reviewed 500 cases in which a lesion was identified histologically and determined if the diagnosis could be made on the first level. Furthermore, we examined 50 additional cases of high-grade dysplasia/carcinoma to determine if the higher-grade component was also present on the first level.

MATERIALS AND METHODS

Cases were retrieved for lesions that could account for a colon polyp clinically, and the first level was examined to determine if lesional tissue was present on the first level. Fifty additional cases of higher-grade lesions were included to ensure higher-grade lesions were present on the first level.

RESULTS

Overall, 497/500 (99.4%) of the non-high-grade lesions were present on the first level, whereas 3/500 (0.6%) required the additional level for diagnosis. All 50 high-grade lesions were present on the first level examined.

DISCUSSION

Many pathology practices routinely order upfront levels on all gastrointestinal biopsies, often generating 2 or 3 slides. Additional slides increase costs, increase the likelihood of laboratory-generated errors, and can waste limited tissue on small biopsies for which ancillary studies may be necessary. Our study showed that a single level is sufficient in the overwhelming majority of cases in which a lesion is identified histologically.

Utility of Additional Tissue Sections in Surgical Pathology

Background. To avoid diagnostic errors such as missed diagnosis and errors in staging tumors due to inadequate tissue sampling, pathologists submit additional sections (AS). Objective. This study assessed frequency, diagnostic yield, distribution, and cost of AS. Method. Among 1542 AS cases, we calculated mean AS per case; fraction of AS that altered diagnosis or stage; AS variation by tissue, malignant versus benign lesions, presence or absence of neoadjuvant therapy, mass, margin, lymph nodes, or other source, resident versus pathologist assistant (PA) dissector; and AS cost per case. Results. Overall 9.2 ± 8.8 AS were collected per case. In only 3.8% (58/1542) of cases AS altered diagnosis or stage. Urinary bladder cases provoked the most AS: 19.5 ± 15.1 per case. Significantly more AS came from malignant versus benign lesions (10.8 ± 9.7 vs 7.6 ± 7.5, P = <.0001) and from specimens treated with neoadjuvant therapy versus malignant lesions not so treated (12.3 ± 9.4 vs 10.3 ± 9.8, P = .02). Lymph nodes were sampled more heavily compared with mass, margin, and other sites combined (11.8 ± 11.4 vs 8.9 ± 8.4, P = .003), but in 78.4% (1209/1542) of cases, AS were from mass. Of diagnosis or stage altering AS cases, two thirds (38/58) were from masses, one fifth (11/58) from lymph nodes, a 10th (6/58) from margins, and a 20th (3/58) from other specimen sites. Resident versus pathologist assistant dissection caused no significant AS difference. AS contributed 40% cost per case. Conclusions. AS per case ranged widely; their diagnostic yield was low; they were highest in urinary bladder specimens, in malignant and particularly neoadjuvant-treated lesions. Although lymph nodes were most heavily sampled, most AS were from masses. Resident dissection did not increase AS and cost of AS was high.

Percutaneous CT-guided needle biopsies of musculoskeletal tumors: a 5-year analysis of non-diagnostic biopsies

OBJECTIVE

To study non-diagnostic CT-guided musculoskeletal biopsies and take steps to minimize them. Specifically we asked: (1) What malignant diagnoses have a higher non-diagnostic rate? (2) What factors of a non-diagnostic biopsy may warrant more aggressive pursuit? (3) Do intra-procedural frozen pathology (FP) or point-of-care (POC) cytology reduce the non-diagnostic biopsy rate?

MATERIALS AND METHODS

This study was IRB-approved and HIPAA-compliant. We retrospectively reviewed 963 consecutive CT-guided musculoskeletal biopsies. We categorized pathology results as malignant, benign, or non-diagnostic and recorded use of FP or POC cytology. Initial biopsy indication, final diagnosis, method of obtaining the final diagnosis of non-diagnostic biopsies, age of the patient, and years of biopsy attending experience were recorded. Groups were compared using Pearson's χ(2) test or Fisher's exact test.

RESULTS

In all, 140 of 963 (15%) biopsies were non-diagnostic. Lymphoma resulted in more non-diagnostic biopsies (P < 0.0001). While 67% of non-diagnostic biopsies yielded benign diagnoses, 33% yielded malignant diagnoses. Patients whose percutaneous biopsy was indicated due to the clinical context without malignancy history almost always generated benign results (96%). Whereas 56% of biopsies whose indication was an imaging finding of a treatable lesion were malignant, 20% of biopsies whose indication was a history of malignancy were malignant. There was no statistically significant difference in the nondiagnostic biopsy rates of pediatric versus adult patients (P = 0.8) and of biopsy attendings with fewer versus more years of experience (P = 0.5). The non-diagnostic rates of biopsies with FP (8%), POC cytology (25%), or neither (24%) were significantly different (P < 0.0001).

CONCLUSION

Lymphoma is the malignant diagnosis most likely to result in a non-diagnostic biopsy. If the clinical and radiologic suspicion for malignancy is high, repeat biopsy is warranted. If the clinical context suggests a benign lesion, a non-diagnostic biopsy may be considered reassuring. Frozen pathology may decrease the non-diagnostic biopsy rate.

Negative colorectal polyp biopsies: the utility of cutting deeper levels

The objectives of this study were to determine the frequency with which deeper levels reveal a lesion in polyp biopsies where no polyp was found on initial sections and to identify features that predict such occult (histologically unapparent) lesions. All initially negative biopsy specimens were accumulated over an 18-month period. Following standard sections, three to ten levels were cut, 50 μm apart. The presence of any lesion, the level at which it was found, the location, number and size of fragments, number of levels obtained, presence of any lymphoid aggregate, endoscopic size and appearance, and bowel preparation quality were recorded. There were 214 specimens, mean patient age 61.4 years (range 27-86 years). Deeper levels revealed a lesion in 52/214 (24.3 %) cases; 76.9 % were tubular adenomas (TA), 21.2 % were hyperplastic polyps, and one was a leiomyoma. All TAs were negative for high-grade dysplasia and malignancy. The mean level at which TAs were found was 1.85 (range 1-9). Male sex (p = 0.021) and right-sided location (p = 0.0075) were statistically significant predictors of an occult TA. As the presence of an adenoma affects screening, pathologists should consider "pursuing" polyps when initial sections reveal no lesion, after ascertaining the incidence of occult lesions in their own practice.