Deep learning system for true- and pseudo-invasion in colorectal polyps.

Over 15 million colonoscopies were performed yearly in North America, during which biopsies were taken for pathological examination to identify abnormalities. Distinguishing between true- and pseudo-invasion in colon polyps is critical in treatment planning. Surgical resection of the colon is often the treatment option for true invasion, whereas observation is recommended for pseudo-invasion. The task of identifying true- vs pseudo-invasion, however, could be highly challenging. There is no specialized software tool for this task, and no well-annotated dataset is available. In our work, we obtained (only) 150 whole-slide images (WSIs) from the London Health Science Centre. We built three deep neural networks representing different magnifications in WSIs, mimicking the workflow of pathologists. We also built an online tool for pathologists to annotate WSIs to train our deep neural networks. Results showed that our novel system classifies tissue types with 95.3% accuracy and differentiates true- and pseudo-invasions with 83.9% accuracy. The system's efficiency is comparable to an expert pathologist. Our system can also be easily adjusted to serve as a confirmatory or screening tool. Our system (available at http://ai4path.ca ) will lead to better, faster patient care and reduced healthcare costs.

Impact of COVID-19-related health care disruptions on pathologic cancer staging during the first pandemic year: a retrospective cohort study from March 2018 to March 2021.

BACKGROUND: The COVID-19 pandemic has created major disruptions in cancer care, with reductions in diagnostic tests and treatments. We evaluated the impact of these health care-related changes on cancer staging by comparing cancers staged before and during the pandemic. METHODS: We performed a retrospective cohort study at London Health Sciences Centre and St. Joseph's Health Care London, London, Ontario, Canada. We evaluated all pathologically staged breast, colorectal, prostate, endometrial and lung cancers (the 5 most common cancers by site, excluding nonmelanoma skin cancer) over a 3-year period (Mar. 15, 2018-Mar. 14, 2021). The pre-COVID-19 group included procedures performed between Mar. 15, 2018, and Mar. 14, 2020, and the COVID-19 group included procedures performed between Mar. 15, 2020, and Mar. 14, 2021. The primary outcome was cancer stage group, based on the pathologic tumour, lymph node, metastasis system. We performed univariate analyses to compare demographic characteristics, pathologic features and cancer stage between the 2 groups. We performed multivariable ordinal regression analyses using the proportional odds model to evaluate the association between stage and timing of staging (before v. during the pandemic). RESULTS: There were 4055 cases across the 5 cancer sites. The average number of breast cancer staging procedures per 30 days increased during the pandemic compared to the yearly average in the pre-COVID-19 period (41.3 v. 39.6), whereas decreases were observed for endometrial cancer (15.9 v. 16.4), colorectal cancer (21.8 v. 24.3), prostate cancer (13.6 v. 18.5) and lung cancer (11.5 v. 15.9). For all cancer sites, there were no statistically significant differences in demographic characteristics, pathologic features or cancer stage between the 2 groups (p > 0.05). In multivariable regression analysis, for all cancer sites, cases staged during the pandemic were not associated with higher stage (breast: odds ratio [OR] 1.071, 95% confidence interval [CI] 0.826-1.388; colorectal: OR 1.201, 95% CI 0.869-1.661; endometrium: OR 0.792, 95% CI 0.495-1.252; prostate: OR 1.171, 95% CI 0.765-1.794; and lung: OR 0.826, 95% CI 0.535-1.262). INTERPRETATION: Cancer cases staged during the first year of the COVID-19 pandemic were not associated with higher stage; this likely reflects the prioritization of cancer procedures during times of reduced capacity. The impact of the pandemic period on staging procedures varied between cancer sites, which may reflect differences in clinical presentation, detection and treatment.

Staying Online in Uncertain Times: A Nationwide Canadian Survey of Pathology Resident Uses of and Adaptations to Online Learning During COVID-19.

CONTEXT.­: Because of restrictions as a result of the COVID-19 pandemic, medical educators rapidly transitioned to an online curriculum for pathology resident education. The benefits and challenges of the shift to online learning, as well as strategies to maximize learning, are yet to be fully elucidated. OBJECTIVE.­: To assess learner perception and satisfaction with the move to online learning. Understanding the benefits of online learning will allow future curricular changes to most effectively incorporate online learning. Understanding the common challenges will allow our current learning strategies to rapidly adapt and ideally mitigate these challenges as online learning is incorporated into medical education. DESIGN.­: This was a survey-based study distributed by email to pathology residents nationwide in Canada in anatomic pathology, general pathology, neuropathology, and hematopathology. Thirty residents participated, from anatomic pathology (n = 23; 76%), from general pathology (n = 5; 16%), and 1 participant each from hematopathology and neuropathology. RESULTS.­: All participants indicated that their program had transitioned to online learning at least in part. The majority of participants (n = 16; 53%) did not feel their pathology education was negatively affected by the transition to online learning; however, a significant minority (n = 6; 20%) felt their education had been negatively affected. Convenience and less intimidation were rated as benefits of online learning. Negative effects included technical issues and decreased engagement; we identified a number of strategies used by programs and pathology residents to mitigate these negative effects. CONCLUSIONS.­: Our survey points to a need to use adaptations and best-practice recommendations to maximize the benefits of online learning moving forward.

Impact of COVID-19 pandemic on neuropathology service: Experience at one Canadian center.

The COVID-19 pandemic has had a significant impact on medical services. Many countries postponed nonemergent procedures to preserve hospital resources for the unprecedented situation. Surgical backlogs caused by the COVID-19 pandemic have been evaluated by different groups. However, the impact of this pandemic on pathology and specifically neuropathology (NP) services has received limited attention. In this study, we reviewed all NP reports of the London Health Sciences Centre from January 2018 (2 years before the pandemic declaration) until the end of the year 2021. Demographic information and pathology details were collected. For tumors, site, histopathology types, and WHO grading were analyzed. In nontumoral specimens, pathological diagnoses were compared in pre- and postpandemic time. The total number of NP samples reached its lowest in April 2020, corresponding to the first Ontario provincial lockdown, and fluctuated throughout the studied period. Among the different types of NP surgical specimens, muscle and epilepsy-related specimens showed a more significant reduction, compared to neoplastic specimens. In 2020, the proportion of tumor specimens from patients older than 40 years of age increased. Similarly, the proportion of high-grade glioma and brain metastasis diagnoses also increased. Lastly, we observed a marked increase in biopsies for temporal arteritis and other inflammatory lesions.

Routine Elastin Staining in Surgically Resected Colorectal Cancer: Impact on Venous Invasion Detection and its Association With Oncologic Outcomes.

Venous invasion (VI) is a powerful yet underreported prognostic factor in colorectal cancer (CRC). Its detection can be improved with an elastin stain. We evaluated the impact of routine elastin staining on VI detection in resected CRC and its relationship with oncologic outcomes. Pathology reports from the year before (n=145) and the year following (n=128) the implementation of routine elastin staining at our institution were reviewed for established prognostic factors, including VI. A second review, using elastin stains, documented the presence/absence, location, number, and size of VI foci. The relationship between VI and oncologic outcomes was evaluated for original and review assessments. VI detection rates increased from 21% to 45% following implementation of routine elastin staining (odds ratio [OR]=3.1; 95% confidence interval [CI]: 1.8-5.3; P<0.0001). The second review revealed a lower VI miss rate postimplementation than preimplementation (22% vs. 48%, respectively; P=0.007); this difference was even greater for extramural VI-positive cases (9% vs. 38%, respectively; P=0.0003). Missed VI cases postimplementation had fewer VI foci per missed case (P=0.02) and a trend towards less extramural VI than those missed preimplementation. VI assessed with an elastin stain was significantly associated with recurrence-free survival (P=0.003), and cancer-specific survival (P=0.01) in contrast to VI assessed on hematoxylin and eosin alone (P=0.053 and 0.1, respectively). The association between VI and hematogenous metastasis was far stronger for elastin-detected VI (OR=11.5; 95% CI: 3.4-37.1; P<0.0001) than for hematoxylin and eosin-detected VI (OR=3.7; 95% CI: 1.4-9.9; P=0.01). Routine elastin staining enhances VI detection and its ability to stratify risk in CRC and should be considered for evaluation of CRC resection specimens.

Evaluating the impact of lymph node resampling on colorectal cancer nodal stage.

AIMS: Nodal staging in colorectal cancer (CRC) informs prognosis and guides adjuvant treatment decisions. A standard minimum of 12 lymph nodes is widely used, with additional sampling being performed as required. However, there are few data on how lymph node resampling in this context has an impact on nodal stage. The aims of this study were to evaluate the effectiveness of resampling in detecting metastases and tumour deposits, and the impact on stage. METHODS AND RESULTS: A retrospective cohort analysis was performed on CRC resections that underwent resampling because of an initial yield of <12 lymph nodes, from 2008 to 2018. Data relating to patient demographics, specimen, malignancy and prosection were collected. Slides were reviewed to quantify nodal metastases and tumour deposits before and after resampling. Among ≥pN1 cases, logistic regression analysis was performed to evaluate factors that predicted the finding of additional metastases and tumour deposits. The cohort comprised 395 cases: resampling identified nodal metastases and/or tumour deposits in 30 (7.6%) cases; nodal upstaging occurred in 20 (5.1%) cases; and eight (2.0%) cases changed from pN0 to ≥pN1. No factors predicted resampling of positive lymph nodes or tumour deposits, and pN upstaging occurred across a variety of cases. A subgroup analysis was performed to assess the impact of resampling on high-risk features in stage II cases (n = 117). There were 33 (8.5%) patients who no longer had any high-risk features after resampling. CONCLUSIONS: Lymph node resampling has an impact on nodal staging and possible treatment decisions in a considerable proportion of patients, and is recommended in all cases with <12 lymph nodes.

Biopsies From Ascending and Descending Colon Are Sufficient for Diagnosis of Microscopic Colitis.

BACKGROUND & AIMS: Lymphocytic and collagenous colitis are types of microscopic colitis (MC) that commonly cause chronic watery diarrhea, but there are no macroscopic features of MC that can be detected during colonoscopy. Endoscopists therefore often collect multiple random colonic biopsies, potentially oversampling, increasing times of colonoscopy and slide review. We sought to identify sites from which biopsies could be taken and analyzed to identify patients with MC with a high level of sensitivity and determine the appropriate number of biopsies to take at these sites. METHODS: We performed a retrospective study using biopsies from 101 consecutive patients with MC (52 cases of collagenous colitis, 42 cases of lymphocytic colitis, 7 combined cases), without comorbidities, from 2017 through 2018. Slides were reviewed, and the proportion of biopsies that were diagnostic of MC were calculated at each biopsy site. RESULTS: The proportions of biopsy fragments from each site of the colon found to be positive for MC were as follows: cecum, 90.0%; ascending colon, 96.9%; hepatic flexure, 77.8%; transverse colon, 95.7%; splenic flexure, 75.0%; descending colon, 85.0%; sigmoid colon, 90.9%; and rectum, 82.2%. For biopsies labeled random, 95.7% were positive for MC. When findings from ascending and descending colon biopsies were combined, 100% of MC cases were detected. CONCLUSIONS: MC can be detected with certainty by analyzing biopsies from the ascending and descending colon. Fewer biopsies than were collected from our cases are sufficient for diagnosis. We propose a Western protocol (taking 2 biopsies from each of the ascending and descending colon) in evaluation of patients for MC.

Challenges with colorectal cancer staging: results of an international study.

Challenges exist with standardized colorectal cancer reporting despite adoption of the American Joint Committee on Cancer-Staging Manual 8th edition. We performed this study to gauge current practice patterns among a diverse group of surgical pathologists. A web-based questionnaire depicting problematic issues and images related to colorectal carcinoma staging was circulated among 118 surgical pathologists and their responses were correlated with their geographic location (North America vs. Europe vs. others), nature of practice (academic vs. community), the sign-out model (gastrointestinal subspecialty vs. general surgical pathology), and years of professional experience. We found that a substantial number of practicing pathologists ignore recommended-staging criteria in specific settings, particularly with respect to assessment of advanced T stage. Tumors that communicated with the serosa through inflammatory foci were staged as pT3 (49%) or pT4a (51%) by nearly equal numbers of pathologists regardless of level of experience, the sign-out model, or geographic location. Only 65% assigned T stage and margin status based on extent of viable tumor in the neoadjuvant setting. One-third of pathologists, particularly those in Europe (p = 0.015), classified acellular mucin deposits as N1 disease when detected in treatment-naive cases. Nearly 50% of pathologists classified isolated tumor cells (i.e., deposits <0.2 mm) in lymph nodes as metastatic disease (i.e., pN1, p = 0.02). Our results suggest that pathologists ignore recommendations that are based on insufficient data and apply individualized criteria when faced with situations that are not addressed in the American Joint Committee on Cancer Staging Manual 8th edition. These variations in practice limit the ability to compare outcome data across different institutions and draw attention to areas that require further study.

Pimping in Residency: The Emotional Roller-Coaster of a Pedagogical Method - A Qualitative Study Using Interviews and Rich Picture Drawings.

Phenomenon: Pimping has become a well-known and distinct form of questioning in medical education, and as a pedagogical method it has both proponents and detractors. Pimping occurs when a teacher (pimper) asks difficult questions of the learner (pimpee), usually in rapid succession. There is a paucity of literature formally studying this technique and its effects on teachers and learners. Our study examines the use of and attitudes toward pimping in a pathology residency program to better understand its perceived value and effectiveness. Approach: Using a qualitative approach, we conducted semistructured interviews with 8 pathology trainees and 9 pathologists. As part of the interview process, we asked participants to draw a rich picture of a pimping encounter. Consistent with this qualitative method, we analyzed data iteratively using constant comparison. Findings: Negative emotions including anxiety and self-doubt dominated among the learners during pimping encounters. For some, these resulted in motivation to study, and for others this was a futile, nonmotivating experience. Most trainees felt that they were being judged during pimping; however, they perceived that the intentions of pimping were not malicious and in their best interests. This was supported by pathologists, who stated that their motivation for pimping was to identify knowledge gaps, thus benefiting the trainee. Insights: Pimping created a dichotomy of emotions within the majority of learners in this study. Negative emotions occurred during pimping encounters; however, following the encounter, pimping was perceived in a more positive light. Recognizing when and how pimping can create negative emotions that may interfere with learning may enable educators to create more consistently meaningful interactions.

CDX2 and Muc2 immunohistochemistry as prognostic markers in stage II colon cancer.

The treatment for colorectal cancer is largely surgical followed by adjuvant chemotherapy in high-risk cases. In patients with stage II cancer, there is no clear benefit for chemotherapy, and the current tools for assessment of risk are inadequate. A recent study identified that colorectal cancer with a gene signature similar to undifferentiated colonic stem cells was associated with a worse outcome. It was later shown that loss of CDX2 detected by immunohistochemistry (IHC) alone resulted in a worse prognosis and that this could be used to predict patients who would benefit from chemotherapy. Having observed that CDX2 expression can be patchy, we elected to validate these prior results for clinical practice using whole-slide IHC. The pathology of all cases was reviewed, and 3 blocks were selected for CDX2 IHC. We also expanded the panel beyond CDX2 to assess whether other markers in the gene signature including CDX1, Muc2, GPX2, and villin could better predict outcome. Among 210 cases, CDX2 expression was diffusely lost in 11% and focally lost in 23% of cases. There was no difference in survival based on CDX2 expression, but Muc2 loss was associated with reduced survival (hazard ratio, 3.32; 95% confidence interval, 1.20 to 9.20). No significant differences in outcome were identified based on CDX1, GPX2, or villin expression. In keeping with this, assessment of The Cancer Genome Atlas gene expression data demonstrated that decreased Muc2 expression was associated with reduced overall survival. Our results with whole-slide IHC are different from the previous studies and caution against the use of CDX2 in isolation as a prognostic marker in clinical practice. We have identified that loss of Muc2 is associated with reduced survival. This supports the use of the colonic differentiation gene expression signature to identify high-risk patients but cautions against the use of any one IHC-based marker in isolation.

Grading of Total Mesorectal Excision Specimens: Assessment of Interrater Agreement.

BACKGROUND: Total mesorectal excision is the standard of care for patients with rectal cancer. Pathological evaluation of the quality of the total mesorectal excision specimen is an important prognostic factor that correlates with local recurrence, but is potentially subjective. OBJECTIVE: This study aimed to determine the degree of variation in grading, both between assessors and between fresh and formalin-fixed specimens. DESIGN: Raters included surgeons, pathologists, pathology residents, pathologists' assistants, and pathologists' assistant trainees. Specimens were assessed by up to 6 raters in the fresh state and by 2 raters postfixation. Four parameters were evaluated: mesorectal bulk, surface regularity, defects, and coning. Interrater agreement was measured using ordinal α-values. SETTING: The study was conducted at a single academic center. MAIN OUTCOME MEASURES: The primary outcome was agreement between individuals when grading total mesorectal excision specimens. RESULTS: A total of 37 total mesorectal excision specimens were assessed. Reliability between all raters for fresh specimens for mesorectal bulk, surface regularity, defects, coning, and overall grade were 0.85, 0.85, 0.92, 0.84, and 0.91. When compared with all raters, pathologists and residents had higher agreement and pathologists and surgeons had lower agreement. Ordinal α-values comparing pathologist and pathologist's assistant agreement for overall grade were similar pre- and postfixation (0.78 vs 0.80), but agreement for assessing defects decreased postfixation. Among pathologists' assistants, agreement was higher when grading specimens postfixation than when grading fresh specimens. LIMITATIONS: Assessment bias may have occurred because of the greater number of pathologists' assistants participating than the number of residents and pathologists. CONCLUSIONS: The results indicate good interrater agreement for the assessment of overall grade, with defects showing the best interrater agreement in fresh specimens. Although total mesorectal excision specimens may be consistently graded postfixation, the assessment of defects postfixation may be less reliable. This study highlights the need for additional knowledge-transfer activities to ensure consistency and accurate grading of total mesorectal excision specimens. See Video Abstract at http://links.lww.com/DCR/A497.

The impact of knowledge transfer on the detection of venous invasion in colorectal cancer.

Venous invasion (VI) is an independent predictor of hematogenous metastasis and mortality in colorectal cancer (CRC) yet remains widely underreported. Its detection may require recognition of subtle morphologic clues, which at times are only unmasked with an elastin stain. This study evaluates the impact of a knowledge transfer initiative (KTI) on VI detection in a "real-world" pathology practice setting. Following participation in an interobserver variability study of VI detection (Kirsch et al, 2013), 12 participants received educational materials highlighting key issues in VI detection. Eighteen months later, participants were invited to submit pathology reports from all CRC resections signed out 18 months prior to and 18 months following the KTI (n = 266 and n = 244, respectively). Nine pathologists participated. Reports were reviewed for VI and other established prognostic factors. Numbers of elastin stains and tumor-containing blocks were also recorded. Comparative analyses were adjusted for baseline differences in tumor, lymph node, and metastasis stage; tumor location; use of neoadjuvant therapy; and number of tumor-containing blocks. VI detection increased significantly post-KTI versus pre-KTI (39.3% versus 18.4%, adjusted odds ratio [OR] 2.86 [1.91-4.28], P < .001). Increased VI detection post-KTI was observed in both stage II (31.8% versus 12.5%, adjusted OR 3.27 [1.45-7.42], P = .004) and stage III CRC (62.4% versus 28.2%, adjusted OR 4.23 [2.37-7.55], P < .001). All pathologists demonstrated increased VI detection post-KTI. Use of elastin stains was significantly higher post-KTI versus pre-KTI (61.5% versus 5.3% of cases respectively, P < .001). This study demonstrates the effectiveness of knowledge transfer in increasing VI detection in routine pathology practice.

Gastric foveolar dysplasia: a survey of reporting habits and diagnostic criteria.

This study aimed to ascertain views, incidence of reporting and diagnostic criteria for gastric foveolar dysplasia. A questionnaire, a post-questionnaire discussion and microscopic assessment of selected cases was conducted by gastrointestinal pathologists to explore the above-stated aims. Fifty-four percent of respondents never or rarely diagnosed gastric foveolar-type dysplasia. The general consensus was that round nuclei, lack of nuclear stratification, presence of inflammation/damage and surface maturation favoured reactive change; while architectural abnormalities/complexity and nuclear enlargement mainly were used to separate low-grade from high-grade foveolar dysplasia. Immunohistochemistry was rarely used to make the diagnosis of dysplasia and was thought not to be of help in routine practice. Inter-observer agreement in grading of dysplasia versus reactive, and the type of dysplasia (foveolar versus adenomatous), was substantial/almost perfect amongst 35.7% and 21.4% of participants, respectively. This reflects low reproducibility in making these diagnoses. In conclusion, foveolar dysplasia was a rarely made diagnosis among 14 gastrointestinal pathologists, there are no uniform criteria for diagnosis and there is poor inter-observer agreement in separating low-grade foveolar dysplasia from reactive gastric mucosa and low-grade adenomatous dysplasia. Greater awareness and agreed criteria will prevent misdiagnosis of low-grade foveolar dysplasia as reactive, and vice versa.

Development and validation of a histological index for UC.

OBJECTIVE: Although the Geboes score (GS) and modified Riley score (MRS) are commonly used to evaluate histological disease activity in UC, their operating properties are unknown. Accordingly, we developed an alternative instrument. DESIGN: Four pathologists scored 48 UC colon biopsies using the GS, MRS and a visual analogue scale global rating. Intra-rater and inter-rater reliability for each index and individual index items were measured using intraclass correlation coefficients (ICCs). Items with high reliability were used to develop the Robarts histopathology index (RHI). The responsiveness/validity of the RHI and multiple histological, endoscopic and clinical outcome measures were evaluated by analyses of change scores, standardised effect size (SES) and Guyatt's responsiveness statistic (GRS) using data from a clinical trial of an effective therapy. RESULTS: Inter-rater ICCs (95% CIs) for the total GS and MRS scores were 0.79 (0.63 to 0.87) and 0.80 (0.69 to 0.87). The correlation estimates between change scores in RHI and change score in GS and MRS were 0.75 (0.67 to 0.82) and 0.84 (0.79 to 0.88), respectively. The SES and GRS estimates for GS, MRS and RHI were: 1.87 (1.54 to 2.20) and 1.23 (0.97 to 1.50), 1.29 (1.02 to 1.56) and 0.88 (0.65 to 1.12), and 1.05 (0.79 to 1.30) and 0.88 (0.64 to 1.12), respectively. CONCLUSIONS: The RHI is a new histopathological index with favourable operating properties.

Loss of the retinoblastoma tumor suppressor correlates with improved outcome in patients with lung adenocarcinoma treated with surgery and chemotherapy.

The retinoblastoma tumor suppressor pathway is frequently inactivated in human cancer, enabling unrestrained proliferation. Most cancers, however, maintain expression of a wild-type (WT) retinoblastoma tumor suppressor protein (pRB). It is generally in a hyperphosphorylated state (ppRB) because of mutations in upstream regulators such as p16 and cyclin D. Hyperphosphorylated ppRB is considered inactive, although data are emerging that suggest it can retain some function. To test the clinical relevance of pRB status, we obtained archival tissue sections from 91 cases of lung adenocarcinoma resected between 2003 and 2008. All cases received platinum doublet chemotherapy, and the median survival was 5.9 years. All cases were assessed for pRB and ppRB using immunohistochemistry and quantified based on intensity of signal and proportion of positive cells. pRB expression was lost in 15% of lung adenocarcinoma cases. In tumors that did not express pRB, the survival rate was significantly improved (hazard ratio, 0.21; 95% confidence interval, 0.06-0.69; P = .01) in comparison to tumors that express pRB. pRB status was found to be an independent predictor of overall survival on multivariate analysis (hazard ratio, 0.22; 95% confidence interval, 0.07-0.73; P = .01) along with increased stage and age. pRB status did not alter baseline levels of apoptotic or proliferative markers in these tumors, but the DNA damage response protein 53BP1 was higher in cancers with high levels of pRB. In summary, loss of pRB expression is associated with improved survival in patients treated with surgical resection and chemotherapy. This may be useful in classifying patients at greatest benefit for aggressive treatment regimes.

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.