Machine Learning Interpretability Methods to Characterize the Importance of Hematologic Biomarkers in Prognosticating Patients with Suspected Infection.

Early detection of sepsis in patients admitted to the emergency department (ED) is an important clinical objective as early identification and treatment can help reduce morbidity and mortality rate of 20% or higher. Hematologic changes during sepsis-associated organ dysfunction are well established and a new biomarker called Monocyte Distribution Width (MDW) has been recently approved by the US Food and Drug Administration for sepsis. However, MDW, which quantifies monocyte activation in sepsis patients, is not a routinely reported parameter and it requires specialized proprietary laboratory equipment. Further, the relative importance of MDW as compared to other routinely available hematologic parameters and vital signs has not been studied, which makes it difficult for resource constrained hospital systems to make informed decisions in this regard. To address this issue, we analyzed data from a cohort of ED patients (n=10,229) admitted to a large regional safety-net hospital in Cleveland, Ohio with suspected infection who later developed poor outcomes associated with sepsis. We developed a new analytical framework consisting of seven data models and an ensemble of high accuracy machine learning (ML) algorithms (accuracy values ranging from 0.83 to 0.90) for the prediction of outcomes more common in sepsis than uncomplicated infection (3-day intensive care unit stay or death). To characterize the contributions of individual hematologic parameters, we applied the Local Interpretable Model-Agnostic Explanation (LIME) and Shapley Additive Value (SHAP) interpretability methods to the high accuracy ML algorithms. The ML interpretability results were consistent in their findings that the value of MDW is grossly attenuated in the presence of other routinely reported hematologic parameters and vital signs data. Further, this study for the first time shows that complete blood count with differential (CBC-DIFF) together with vital signs data can be used as a substitute for MDW in high accuracy ML algorithms to screen for poor outcomes associated with sepsis.

Sclerosing Angiomatoid Nodular Transformation of the Spleen: A Report of Rare Case and Literature Review.

Sclerosing angiomatoid nodular transformation (SANT) is a benign vascular lesion of the spleen with uncertain etiology. It predominantly affects women between the ages of 30 and 60 years. Clinically, it is asymptomatic or can cause abdominal pain, but usually discovered incidentally on imaging, which can identify a mass but may not provide a definitive diagnosis. In uncertain vascular lesions, there is always a risk of spontaneous rupture of large vessels and the potential for spreading malignancy. Hence, the final diagnosis is rendered on microscopy after splenectomy. A middle-aged female came to the clinic complaining of abdominal pain. Radiology showed a solid splenic mass and the patient underwent splenectomy. Gross examination showed a 3 cm white firm mass with focal hemorrhage. Microscopy revealed multiple nodules of variable sizes surrounded by fibrosclerotic stroma. The nodules showed round to slit-like vascular spaces with numerous red blood cells. The internodular stroma consisted of dense fibrous tissue with scattered plump myofibroblasts and lymphoplasmacytic inflammatory cells. These distinctive features lead to the diagnosis of SANT. SANT possesses characteristic histologic features with distinctive immunohistochemistry (IHC). IHC reveals three different types of vessels within the nodules as follows: (1) small veins (CD34-, CD31+, CD8-), (2) sinusoids (CD34-, CD31+, CD8+), and (3) capillaries (CD34+, CD31+, CD8-). All three types of vessels are negative for CD21/CD35 and CD68. Hemangioma and littoral cell angioma are two frequent vascular tumors in the spleen that should be considered differential diagnoses. Both lesions lack the microscopic features of SANT and have only a single type of vessel. The vessels in hemangioma are (CD31+, CD34+, CD8-), while in littoral cell angioma they are (CD31+, CD34-, CD8-, CD21+, CD68+). There are no specific clinical or radiologic findings for SANT. It is important to recognize these characteristic features and to differentiate them from other benign and malignant lesions, such as angiosarcoma. A thorough histopathologic examination and IHC are helpful in making the correct diagnosis.

Co-occurrence of myeloid neoplasm and plasma cell neoplasm.

Co-occurrence of myelodysplastic syndrome (MDS) and plasma cell neoplasm in patients with no history of chemo and/or radiotherapy is rarely reported. Herein, we report a case of a female in her seventieth decade of life who was referred to the hospital for pancytopenia. The patient was asymptomatic and was doing well overall. Serum protein electrophoresis was remarkable for a lambda-restricted monoclonal protein (IgG) estimated at 1.8g/dL. Immunoglobulin G serum level was also elevated, and serum Kappa/Lambda free light chain ratio was decreased. At that time, a bone marrow biopsy showed myelodysplastic syndrome with excess blasts-2 (MDS-EB2) and a monoclonal plasma cell proliferation. Some studies have shown that patients with plasma cell neoplasm could be associated with an increased risk of developing MDS compared to the general population. Based on reviewing the literature, to our knowledge, the pathological mechanism of the co-occurrence of both diseases is not yet clear.

Co-occurrence of myeloid neoplasm and plasma cell neoplasm

ABSTRACT Co-occurrence of myelodysplastic syndrome (MDS) and plasma cell neoplasm in patients with no history of chemo and/or radiotherapy is rarely reported. Herein, we report a case of a female in her seventieth decade of life who was referred to the hospital for pancytopenia. The patient was asymptomatic and was doing well overall. Serum protein electrophoresis was remarkable for a lambda-restricted monoclonal protein (IgG) estimated at 1.8g/dL. Immunoglobulin G serum level was also elevated, and serum Kappa/Lambda free light chain ratio was decreased. At that time, a bone marrow biopsy showed myelodysplastic syndrome with excess blasts-2 (MDS-EB2) and a monoclonal plasma cell proliferation. Some studies have shown that patients with plasma cell neoplasm could be associated with an increased risk of developing MDS compared to the general population. Based on reviewing the literature, to our knowledge, the pathological mechanism of the co-occurrence of both diseases is not yet clear.

Pancreatic lymphoma: A cytologic diagnosis challenge.

Very rarely lymphoma primarily or secondarily involves the pancreas. Involvement of the pancreatic parenchyma with lymphoma clinically may mimic pancreatic ductal adenocarcinoma (PDA) and other mass-forming pancreatic lesions. Endoscopic ultrasound fine needle aspiration (EUS-FNA) is the first step in the diagnostic pathway of managing these patients by providing a cytology specimen. Cytologically, lymphoma of pancreas can be misdiagnosed for a wide variety of pancreatic neoplastic and non-neoplastic lesions. Cytological differential diagnosis includes well-differentiated adenocarcinoma, acinar cell carcinoma, well differentiated neuroendocrine tumor, and autoimmune pancreatitis. Gastroenterologist's skills in providing adequate sample for preparing smears, cell blocks and/or performing flow cytometry, and also cytopathologist's skills in detecting atypical lymphocytic population are crucial factors. Although cytology examination has limitations to subclassify lymphoma, it plays a key role to redirect clinicians into the right patient-care pathway. In this article, we present two cases of pancreatic lymphoma with emphasis on the discriminating cytomorphological features, and we also review literatures with reports of primary pancreatic lymphoma (PPL) to better understand the characteristics of this rare lesion.

Lipid-rich variant of pancreatic endocrine neoplasms: a case report.

BACKGROUND: Pancreatic endocrine neoplasms (PENs) are a well-defined and well-characterized group of tumors. We report a rare variant of PENs called a lipid-rich variant and elaborate on its cytologic features with reference to fine needle aspiration (FNA). CASE: The patient was a 74-year-old man with a 2.1-cm mass in the body of the pancreas. The FNA showed numerous cells with vacuolated, foamy cytoplasm, mostly arranged individually but also in some apparently cohesive clusters. The nuclei were quite uniform. Many of the cells had the appearance of histiocytes. The cells in the cell block were immunoreactive (positive) for cytokeratin AE1/AE3, synaptophysin, and chromogranin A. Some admixed histiocytes stained for CD68. A diagnosis of PEN was made, with a suspicion for lipid-rich variant. An elective resection of the tumor was performed, which confirmed the diagnosis of lipid-rich variant of PENs. Lipid-rich variant of PENs can present a diagnostic dilemma on FNA. CONCLUSION: Its mimickers include adrenal cortical carcinoma, metastatic clear-cell renal cell carcinoma, clear-cell PEN, foamy gland pattern of pancreatic ductal carcinoma, solid pseudopapillary tumor, and acinar cell carcinoma. The distinguishing morphologic and immunohistochemical features of each are described.

Prostatic adenocarcinoma in colorectal biopsy: clinical and pathologic features.

Locally advanced prostate carcinoma (PCa) can involve adjacent colorectal tissue. In such cases, accurate assignment of primary source may prove difficult. Given the difference in treatment and prognosis between PCa and colorectal carcinoma (CRCa), an accurate diagnosis is crucial. Surgical pathology files were searched for all cases with a diagnosis of PCa on colorectal biopsy between 1987 and 2006. Histologic and clinical data were available in 23 of the 30 found cases. The diagnosis of PCa was made for the first time at the time of the colorectal presentation in 4 men. Three of the 4 underwent colectomy. Among the remaining 19 men with a previously documented diagnosis of PCa, the overall clinical and endoscopic impression still favored a second primary CRCa in 12 patients. The latter was due in part to the chronologically distant prior PCa diagnosis (mean, 6.9 years; range, 2-18 years). None of the colorectal biopsies demonstrated carcinoma in situ or dysplastic colonic mucosa. PCa architecture in the colonic biopsies was that of Gleason score of 9 to 10 in 20 cases and Gleason score of 8 in the remaining 3. Cribriform or microacinar architecture typical of PCa was seen in 6 cases. Immunostains were positive for prostate-specific antigen, P501S (prostein), prostate-specific membrane antigen, and prostate-specific acid phosphatase, in 16 of 20, 9 of 11, 10 of 11, and 10 of 20 cases, respectively. Three cases lacked immunohistochemical evidence of prostatic differentiation. Caudal-type homebox transcription factor 2 (CDX2) and beta-catenin were negative in all tested cases (0/11 for both). Although relatively rare, initial presentation of PCa as a rectal lesion may lead to an erroneous clinical and or histologic impression of CRCa. Because history of prostate cancer is often remote, clinical findings may be misleading. Pathologists should consider the possibility of prostatic origin in poorly differentiated carcinoma encountered on colorectal biopsy when features such as lack of an in situ component, extrinsic pattern of involvement, microacinar or solid architecture, and/or prominent nucleoli, are noted, especially in the absence of nuclear pleomorphism and mitotic activity. Immunohistochemical studies could help establish the diagnosis.