Chronic enteropathy (CE) is the most common gastrointestinal disorder of older cats and has been increasing in prevalence. The most common constituents of the disorder include lymphocytic-plasmacytic enteropathy (LPE) and small cell lymphoma (SCL). Typical diagnostic evaluation includes collection of intestinal biopsy samples (either endoscopically or surgically) and histopathologic evaluation of tissues. Several techniques exist to aid in the differentiation of LPE and SCL, including immunohistochemistry (IHC) with stains for T- and B- cells, as well as PCR-based clonality assays. However, there exists significant variation of sensitivities and specificities among techniques, and there is currently no gold standard for the diagnosis and differentiation of LPE from SCL in cats.
Histology-guided mass spectrometry (HGMS) profiling applies matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI), and employs histopathologic annotations for targeted analysis of endogenous molecules from specific cell subpopulations. The technology utilizes a machine learning algorithm to assess the distribution of targeted molecules within a tissue sample, such as proteins, lipids, or metabolites, to differentiate molecular signatures and specific disease states. HGMS technology has been demonstrated in animal models, and successfully applied in human medicine to differentiate benign nevi from malignant melanoma. Similar to melanoma in humans, the differentiation of LPE from SCL in cats can be challenging.
The purpose of the study was to investigate HGMS as a possible new method for the differentiation of LPE from SCL. The authors also sought to compare the performance of PCR for antigen receptor rearrangements (PARR) – a PCR-based clonality assay technique – and HGMS for the differentiation of LPE and SCL in cats. The study was designed as a retrospective clinicopathologic study.
Duodenal tissue samples were selected from a database at the Texas A&M University Gastrointestinal Laboratory from cats with a history CE. A total of 121 cases were first evaluated by a panel of 5-7 veterinary experts (including internists, pathologists and oncologists). Cases that received unanimous consensus diagnoses by the experts of either LPE of SCL were included in the HGMS cohort of the study, and cases that were unanimously diagnosed were eligible for use in the algorithm development cohort of the study with HGMS profiling data. Cases where no consensus could be reached, or cases that had diagnoses other than LPE or SCL (such as normal tissue, large cell lymphoma, eosinophilic inflammation, neutrophilic inflammation) were excluded from further analysis.
Overall, 93 cases received a consensus diagnosis of either LPE (41) or SCL (52). The median age of included cats was 11 years (1-17 years), and breeds included domestic (n=71), Maine Coon (n=7), Bengal (n=3), unknown breed (n=3), Siamese (n=2 ), Ragdoll (n=2), and one each of Burmese, Norwegian Forest Cat, American, Russian Blue, and Persian. The majority of samples were obtained endoscopically (91/93), and the remaining were obtained surgically (2/93). No significant difference was identified among the demographic characteristics, age, and sex between the training cohort set and the validation cohort. Based on panel evaluation, 39 cases were selected for the training set (LPE = 19, SCL = 20). Overall, 54 cases were assigned to the validation set (LPE = 26, SCL =27).
Following panel evaluation, HGMS was applied to samples. This involved sample preparation by antigen retrieval, trypsin digestion and MALDI matrix application. Anatomic pathologists then marked samples with 50 um-diameter annotations over areas of lymphocyte cell subpopulations on stained images. The images were then overlaid with the physical samples, and the slide and annotated images were merged to teach mass spectrometer locations for analysis. During sample analysis, proteomic data were acquired, then analysed for algorithm development, whereby machine learning was used to generate a classification algorithm to classify samples as either LPE or SCL. HGMS was performed on an initial training set of samples, followed by a validation set. For the validation set, the sensitivity and specificity of HGMS relative to consensus panel diagnoses were 86.7% and 91.7%, respectively. The relative accuracy of HGMS was determined to be 88.9%.
Lastly, PARR analysis was performed on the validation set, where 34/54 samples were found to be clonal, 8/54 polyclonal, 3/54 oligoclonal, and 7/54 pseudoclonal. Overall sensitivity and specificity of PARR relative to consensus panel diagnoses was 85.7% and 33%, respectively. The relative accuracy of PARR was calculated to be 61.5%. The figure below shows plots of positive (A) and negative (B) predictive values for HGMS and PARR as functions of different disease prevalence.
Several potential limitations of the present study exist, including its retrospective nature, which prevented standardization of the biopsy sample collection methods. Follow-up information was not available for all cases. Moreover, removing samples from analysis that lacked consensus diagnoses may have biased samples, and HGMS and annotations were limited to lamina propria of samples from the upper intestinal tract, and thus data from this study can only be applied to those locations. Lastly, the authors acknowledged that expert opinion (such as by panel consensus) does not equal evidence. However, the authors noted that a multidisciplinary approach is commonly used in human medicine, and is considered good clinical practice, especially for diseases where consensus diagnosis may be challenging (such as in melanoma in humans, or CE in cats).
Despite these limitations, this study served to demonstrate that HGMS may be a suitable method for the detection of SCL in cats, and that in a cohort of cats, HGMS showed more favourable results compared to PARR. The authors emphasize that results of any molecular clonality studies should always be interpreted in the context of the clinical, morphological, and immunophenotypic diagnosis, and in collaboration with various experts in the field of study. (HM)
Marsilio S, Ackermann MR, Lidbury JA, Suchodolski JS, Steiner JM. Results of histopathology, immunohistochemistry, and molecular clonality testing of small intestinal biopsy specimens from clinically healthy client-owned cats. J Vet Intern Med. 2019;33: 551–558.