Diagnostic Concordance of Pathological Methods and Reports of Hematopathologists Compared to Local Nonspecialized Pathologists in the Diagnosis of Lymphoma in Mexico

Background: Distinct diagnostic assays and algorithms are employed by contemporary pathologists when seeking to identify lymphoma. There is a paucity of data on the degree of concordance between pathologists diagnosing lymphoma and its subtypes in different institutional settings in Latin America. The objective of this study was to assess the concordance between lymphoma diagnoses made via tissue biopsy by local pathologists and after review of these specimens by more specialized hematopathologists.

Methods: This prospective, noninterventional and multicenter study was conducted at seven sites in Mexico from January 2017 to October 2017. Eligible biopsy samples were from patients with a previous diagnosis of lymphoma on lymph node biopsy or a diagnosis of extranodal lymphoma, with adequate tissue preservation and adequate amount of tissue for the review analysis. Patients receiving either chemotherapy or corticosteroids before sampling of tissue biopsies were excluded. Seven sites representing local pathologists and three hematopathologists participated in the study. The same biopsy tissues reviewed by the local pathologists were also sent to the hematopathologists participating in this study. Physicians in charge of patients' treatment decided whether to make any changes to a patient's treatment decision when comparing diagnosis results from the local pathologists and hematopathologists. The concordance in diagnosis results were classified into 3 categories: diagnostic agreement (i.e. the local pathologist and hematopathologist diagnoses concurred), minor discrepancy (i.e. there was a difference in diagnosis but it didn't change the treatment decision) and major discrepancy (i.e. there was a difference in diagnosis and it changed the treatment decision based on guidelines from the National Comprehensive Cancer Network).

Results: Of 111 samples received, 105 met eligibility criteria for full review by hematopathologists and were included for full analysis. The median (range) patient age was 53 (16-94) years. More specimens were obtained from women (n = 57; 54.3%). A total of 53 (50.5%) patients were recruited from private institutions and the remaining (52; 49.5%) from public institutions. Within the 105 biopsies, a total of 89 cases were diagnosed as lymphoma by hematopathologists, including non-hodgkin's lymphoma (NHL; n = 72; 68.6%) and hodgkin's lymphoma (HL; n = 17; 16.2%). The most common subtype of NHL diagnosed was diffuse large B cell lymphoma (DLBCL), with a total of 32 cases. Diagnostic agreement was observed in 23 (21.9%) biopsies, minor discrepancies in 32 (30.5%) biopsies and major discrepancies in 50 (47.6%) biopsies, indicating treatment decisions changed in 47.6% of the total cases. Subtypes of lymphoma that local pathologists and hematopathologists most commonly found in diagnostic agreements were DLBCL (8/23; 34.8%) and HL (6/23; 26.1%), potentially due to their higher prevalence in the overall studied samples. Three types of error from the local pathologists were observed in major discrepancies, including ambiguity or lack of full diagnosis (27/50; 54%), a change from malignant to benign lesion (11/50; 22%) and a change of the type of neoplasm (12/50; 24%). Hematopathologists reported more immunohistochemical disease markers per tissue specimen: a median of 8.7 and a mode of 8 compared to a median of 5 and a mode of 0 for local pathologists. The diagnostic concordance varied across the seven study sites; the rate of major discrepancies ranged from 0% to 100% and the rate of diagnostic agreement ranged from 0% to 81.8%. The local pathologist from only one site received formal training in hematopathology and reported no major discrepancies. When excluding results from this site, the diagnostic agreement was observed in 14 (14.9%) biopsies, minor discrepancies in 30 (31.9%) biopsies and major discrepancies in 50 (53.2%) biopsies.

Conclusions: This study showed that physicians from the seven study sites in Mexico changed their original treatment decisions that were initially based on local pathologist's diagnosis in nearly one-half (47.6%) cases after they reviewed the hematopathologist's diagnosis. In addition, there was a wide variation in the percentage of diagnostic agreements and discrepancies among different study sites, where sites with more experienced pathologists demonstrated a lower rate of diagnosis discrepancies in the diagnosis of lymphoma.