Value-based flow testing of chronic lymphoproliferative disorders: a quality improvement project to develop an algorithm to streamline testing and reduce costs

Artificial Intelligence for Clinical Flow Cytometry

In this review, the authors discuss the fundamental principles of machine learning. They explore recent studies and approaches in implementing machine learning into flow cytometry workflows. These applications are promising but not without their shortcomings. Explainability may be the biggest barrier to adoption, as they contain "black boxes" in which a complex network of mathematical processes learns features of data that are not translatable into real language. The authors discuss the current limitations of machine learning models and the possibility that, without a multiinstitutional development process, these applications could have poor generalizability. They also discuss widespread deployment of augmented decision-making.

Flow Cytometric, Morphologic, and Laboratory Comparative Study in Patients With Leukocytosis and Cytopenia

OBJECTIVES

We wanted to evaluate the effectiveness of flow cytometry immunophenotyping (FCI) as a screening test for patients with leukocytosis and cytopenia.

METHODS

We identified 320 patients during August 2016 to December 2016 and evaluated FCI and morphology of peripheral blood smears (PBSs).

RESULTS

The most common indications for FCI included history of hematologic malignancy (HHM, n = 126), leukocytosis (n = 80), and cytopenia (n = 53). Positive FCI rate was low with a range of 4.4% to 12.5% in patients with absolute neutrophilia regardless of HHM, if cases with circulating blasts were excluded. Patients with absolute lymphocytosis had a 93% positive FCI rate. Patients with HHM and pancytopenia showed a higher incidence of positive FCI findings than patients without HHM and with isolated cytopenia. PBS morphology correlated strongly with FCI (P = .0001).

CONCLUSION

PBS evaluation is an accurate and cost-effective screening test. FCI for patients with mature neutrophilia and isolated cytopenia has a very low yield.

Machine Learning Models Improve the Diagnostic Yield of Peripheral Blood Flow Cytometry

OBJECTIVES

Peripheral blood flow cytometry (PBFC) is useful for evaluating circulating hematologic malignancies (HM) but has limited diagnostic value for screening. We used machine learning to evaluate whether clinical history and CBC/differential parameters could improve PBFC utilization.

METHODS

PBFC cases with concurrent/recent CBC/differential were split into training (n = 626) and test (n = 159) cohorts. We classified PBFC results with abnormal blast/lymphoid populations as positive and used two models to predict results.

RESULTS

Positive PBFC results were seen in 58% and 21% of training cases with and without prior HM (P < .001). % neutrophils, absolute lymphocyte count, and % blasts/other cells differed significantly between positive and negative PBFC groups (areas under the curve [AUC] > 0.7). Among test cases, a decision tree model achieved 98% sensitivity and 65% specificity (AUC = 0.906). A logistic regression model achieved 100% sensitivity and 54% specificity (AUC = 0.919).

CONCLUSIONS

We outline machine learning-based triaging strategies to decrease unnecessary utilization of PBFC by 35% to 40%.

Maximising yield of peripheral blood flow cytometry for chronic lymphoproliferative disorders

INTRODUCTION

Flow cytometry is used in the diagnosis of haematological diseases including chronic lymphoproliferative disorders. This audit aims to ascertain the real-world indications for peripheral blood (PB) flow cytometry and which of these are associated with higher diagnostic yields.

METHODS

All PB flow cytometry requests for chronic lymphoproliferative disorders from 1 January 2014 to 31 December 2014 were identified using the laboratory information system. Data including patient demographics, specialty of requestor, lymphocyte count and blood film report (if available), indications for tests and subsequent diagnosis were collected.

RESULTS

A total of 185 requests with median patient age of 60 years were analysed. The main requestor was the Haematology Unit (n = 109; 58.9%) although the diagnostic yield of their requests was not significantly better than other units combined (16.5% vs 13.2%, P = .49). Factors that significantly improved the diagnostic yield of testing were older age, the presence of atypical lymphocytes on the blood film and lymphocytosis (P < .01). Constitutional symptoms and cytopenias were not found to influence the diagnostic yield.

CONCLUSION

PB flow cytometry is a useful tool when used in the appropriate clinical setting. Rationalisation of testing is important to reduce the futility of testing and unnecessary health costs.

Cost-effective approach to the diagnostic workup of B cell lymphoproliferative disorders via optimal integration of flow cytometric data

INTRODUCTION

The workup of lymphoproliferative disorders (LPDs) involves the combined use of flow cytometry (FC) and immunohistochemistry (IHC). This often results in duplicate immunophenotypic testing and adds costs that may not be eligible for reimbursement based on the Medicare National Correct Coding Initiative. We aimed to establish a cost-effective diagnostic algorithm based on initial FC categorization to reduce repetitive immunophenotyping.

METHODS

We retrospectively reviewed 242 cases of suspected LPDs with concurrent FC and IHC testing over a 12-month period. We correlated FC with surgical diagnoses and evaluated the frequency of repeat IHC testing.

RESULTS

Repetitive immunophenotyping was common; overall, 85% of cases had at least one marker repeated. Concordant cases were significantly less likely to have markers repeated than discordant cases. Of concordant B cell malignancies, 57% represented recurrent disease; however, repeat marker usage was not decreased as compared to new diagnoses. The most frequently repeated markers were CD3, CD5, CD10, and CD20.

CONCLUSIONS

We propose that in concordant cases, CD5 and CD10 should not be repeated by IHC; this would decrease the use of these markers by 80% and 76%, respectively. We developed an algorithmic approach to IHC usage that has improved incorporation of FC data at our institution and may reduce healthcare costs.

Diagnosis of chronic lymphoproliferative disorders by flow cytometry using four-color combinations for immunophenotyping: A proposal of the brazilian group of flow cytometry (GBCFLUX)

BACKGROUND

Multiparametric flow cytometry (MFC) is a powerful tool for the diagnosis of hematological malignancies and has been useful for the classification of chronic lymphoproliferative disorders (CLPD) according to the WHO criteria. Following the purposes of the Brazilian Group of Flow Cytometry (GBCFLUX), the aim of this report was to standardize the minimum requirements to achieve an accurate diagnosis in CLPDs, considering the different economic possibilities of the laboratories in our country. Most laboratories in Brazil work with 4-fluorescence flow cytometers, which is why the GBCFLUX CLPD Committee has proposed 4-color monoclonal antibody (MoAb) panels.

METHODS/RESULTS

Panels for screening and diagnosis in B, T and NK lymphoproliferative disorders were developed based on the normal differentiation pathways of these cells and the most frequent phenotypic aberrations. Important markers for prognosis and for minimal residual disease (MRD) evaluation were also included. The MoAb panels presented here were designed based on the diagnostic expertise of the participating laboratories and an extensive literature review.

CONCLUSION

The 4-color panels presented to aid in the diagnosis of lymphoproliferative neoplasms by GBCFLUX aim to provide clinical laboratories with a systematic, step-wise, cost-effective, and reproducible approach to obtain an accurate immunophenotypic diagnosis of the most frequent of these disorders. © 2016 International Clinical Cytometry Society.

Laboratory Test Utilization Management: General Principles and Applications in Hematopathology

As the cost of health care continues to rise and reimbursement rates decrease, there is a growing demand and need to cut overall costs, enhance quality of services, and maintain as a top priority the needs and safety of the patient. In this article, we provide an introduction to test utilization and outline a general approach to creating an efficient, cost-effective test utilization strategy. We also present and discuss 2 test utilization algorithms that are evidence-based and may be of clinical utility as we move toward the future of doing the necessary tests at the right time.