Dataset Information

Automated diagnostic support system with deep learning algorithms for distinction of Philadelphia chromosome-negative myeloproliferative neoplasms using peripheral blood specimen.

ABSTRACT: Philadelphia chromosome-negative myeloproliferative neoplasms (Ph-negative MPNs) such as polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis are characterized by abnormal proliferation of mature bone marrow cell lineages. Since various non-hematologic disorders can also cause leukocytosis, thrombocytosis and polycythemia, the detection of abnormal peripheral blood cells is essential for the diagnostic screening of Ph-negative MPNs. We sought to develop an automated diagnostic support system of Ph-negative MPNs. Our strategy was to combine the complete blood cell count and research parameters obtained by an automated hematology analyzer (Sysmex XN-9000) with morphological parameters that were extracted using a convolutional neural network deep learning system equipped with an Extreme Gradient Boosting (XGBoost)-based decision-making algorithm. The developed system showed promising performance in the differentiation of PV, ET, and MF with high accuracy when compared with those of the human diagnoses, namely:?>?90% sensitivity and?>?90% specificity. The calculated area under the curve of the ROC curves were 0.990, 0.967, and 0.974 for PV, ET, MF, respectively. This study is a step toward establishing a universal automated diagnostic system for all types of hematology disorders.

SUBMITTER: Kimura K

PROVIDER: S-EPMC7873208 | biostudies-literature | 2021 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automated diagnostic support system with deep learning algorithms for distinction of Philadelphia chromosome-negative myeloproliferative neoplasms using peripheral blood specimen.

Kimura Konobu K Ai Tomohiko T Horiuchi Yuki Y Matsuzaki Akihiko A Nishibe Kumiko K Marutani Setsuko S Saito Kaori K Kaniyu Kimiko K Takehara Ikki I Uchihashi Kinya K Ohsaka Akimichi A Tabe Yoko Y

Scientific reports 20210209 1

Philadelphia chromosome-negative myeloproliferative neoplasms (Ph-negative MPNs) such as polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis are characterized by abnormal proliferation of mature bone marrow cell lineages. Since various non-hematologic disorders can also cause leukocytosis, thrombocytosis and polycythemia, the detection of abnormal peripheral blood cells is essential for the diagnostic screening of Ph-negative MPNs. We sought to develop an automated ...[more]

PMID: 33564094

Similar Datasets

Project description:This document updates the recommendations on the management of Philadelphia chromosome-negative myeloproliferative neoplasms (Ph-neg MPNs) published in 2011 by the European LeukemiaNet (ELN) consortium. Recommendations were produced by multiple-step formalized procedures of group discussion. A critical appraisal of evidence by using Grades of Recommendation, Assessment, Development and Evaluation (GRADE) methodology was performed in the areas where at least one randomized clinical trial was published. Seven randomized controlled trials provided the evidence base; earlier phase trials also informed recommendation development. Key differences from the 2011 diagnostic recommendations included: lower threshold values for hemoglobin and hematocrit and bone marrow examination for diagnosis of polycythemia vera (PV), according to the revised WHO criteria; the search for complementary clonal markers, such as ASXL1, EZH2, IDH1/IDH2, and SRSF2 for the diagnosis of myelofibrosis (MF) in patients who test negative for JAK2V617, CALR or MPL driver mutations. Regarding key differences of therapy recommendations, both recombinant interferon alpha and the JAK1/JAK2 inhibitor ruxolitinib are recommended as second-line therapies for PV patients who are intolerant or have inadequate response to hydroxyurea. Ruxolitinib is recommended as first-line approach for MF-associated splenomegaly in patients with intermediate-2 or high-risk disease; in case of intermediate-1 disease, ruxolitinib is recommended in highly symptomatic splenomegaly. Allogeneic stem cell transplantation is recommended for transplant-eligible MF patients with high or intermediate-2 risk score. Allogeneic stem cell transplantation is also recommended for transplant-eligible MF patients with intermediate-1 risk score who present with either refractory, transfusion-dependent anemia, blasts in peripheral blood?>?2%, adverse cytogenetics, or high-risk mutations. In these situations, the transplant procedure should be performed in a controlled setting.

Project description:Background: Inflammation might play a critical role in the pathogenesis and progression of Philadelphia-negative myeloproliferative neoplasms (Ph-MPNs) with elevated inflammatory cytokines in peripheral blood (PB). However, the inflammatory status inside the bone marrow (BM), which is the place of malignancy origin and important microenvironment of neoplasm evolution, has not yet been elucidated. Methods: Inflammatory cytokine profiles in PB and BM of 24 Ph-MPNs patients were measured by a multiplex quantitative inflammation array. Cytokines that correlated between PB and BM were selected and then validated by ELISA in a separate cohort of 52 MPN patients. Furthermore, a panel of cytokines was identified and examined for potential application as non-invasive markers for the diagnosis and prediction of fibrosis progress of MPN subtypes. Results: The levels of G-CSF, I-309, IL-1β, IL-1ra, IL-12p40, IL-15, IL-16, M-CSF, MIG, PDGF-BB, and TIMP-1 in BM supernatants were significantly higher than those in PB (all p < 0.05). Linear correlations between BM and PB levels were found in 13 cytokines, including BLC, Eotaxin-2, I-309, sICAM-1, IL-15, M-CSF, MIP-1α, MIP-1δ, RANTES, TIMP-1, TIMP-2, sTNFRI, and sTNFRII (all R > 0.4 and p < 0.05). Levels of BLC, Eotaxin-2, M-CSF, and TIMP-1 in PB were significantly different from those in health controls (all p < 0.05). In PB, levels of TIMP-1 and Eotaxin-2 in essential thrombocythemia (ET) group were significantly lower than those in groups of prefibrotic primary myelofibrosis (pre-PMF) [TIMP-1: 685.2 (322.2-1,229) ng/ml vs. 1,369 (1,175-1,497) ng/ml, p = 0.0221; Eotaxin-2: 531.4 (317.9-756.6) pg/ml vs. 942.4 (699.3-1,474) pg/ml, p = 0.0393] and primary myelofibrosis (PMF) [TIMP-1: 685.2 (322.2-1229) ng/ml vs. 1,365 (1,115-1,681) ng/ml, p = 0.0043; Eotaxin-2: 531.4 (317.9-756.6) pg/ml vs. 1,010 (818-1,556) pg/ml, p = 0.0030]. The level of TIMP-1 in myelofibrosis (MF) >1 group was significantly higher than that in MF ≤ 1 group. Conclusion: Abnormal inflammatory status is present in MPN, especially in its BM microenvironment. Consistency between PB and BM levels was found in multiple inflammatory cytokines. Circulating cytokine levels of BLC, M-CSF, Eotaxin-2, and TIMP-1 reflected inflammation inside BM niche, suggesting potential diagnostic value for MPN subtypes and prognostic value for fibrosis progression.

Dataset Information

Automated diagnostic support system with deep learning algorithms for distinction of Philadelphia chromosome-negative myeloproliferative neoplasms using peripheral blood specimen.

Publications

Automated diagnostic support system with deep learning algorithms for distinction of Philadelphia chromosome-negative myeloproliferative neoplasms using peripheral blood specimen.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets