Trained hematologists are often only available at tertiary-referral hospitals. The symptoms of leukemia, (e.g. fever) are common and can also be found in other benign conditions like viral infection. Thus, it is not surprising that these patients go initially to primary care general centers and attended by practitioners who are usually unskilled in the interpretation of peripheral blood film. Hence, a computer-based system that can clearly discern lymphoblast cells with accuracy would be useful for the initial screening of peripheral blood film of patients with leukemia symptoms. Subsequently, these patients can be referred to the appropriate institutions for further evaluation and management. Early detection of the presence of leukemic cells in the peripheral blood and the subsequent possibility of a prompt treatment are essential for the patient survival. The morphological differentiation among different types of abnormal lymphoid cells and among blast cells in peripheral blood is a difficult task, which requires high experience and skill. Objective values do not exist to define cytological variables. Moreover, subtle morphological characteristics exist that are exhibited by some lymphoid and leukemic cells, which are shared with reactive lymphoid cells. associated to viral (not malign) infections. This sometimes results in doubts on the correct classification in the daily practice. In addition, immunophenotyping shows some overlap among various lymphoid disorders. The ultimate objective of this project is the development of an automatic recognition system, based on digital image processing and pattern recognition, able to supply an accurate classification among a large group of morphological abnormalities in the leucocytes circulating in the peripheral blood, which correspond to specific lymphoid neoplasms and acute leukemias.The project development is organized in the following activities: (1) lymphoid cell and blast cell images dataset organization; (2) development of the automatic segmentation of the regions of interest (complete cell, nucleus and cell outer region); 3) extraction of a wide spectrum of morphological features; (4) automatic recognition of the following four cell types: normal, reactive and abnormal lymphocytes, as well as blast cells, which will be helpful for a first separation between lymphomas and leukemias, since these pathologies have different treatment and prognosis; (5) automatic recognition of abnormal lymphoid cell types in a specific lymphoma; (6) automatic recognition between myeloid and lymphoid blasts, which is important since they require different treatment; 7) integration of the recognition system and proof of concept with images from patients of three hospitals; and (8) dissemination and transfer of results. The outcome of the project will be an innovative support tool for quick initial diagnosis of hematological malignancies, highly valuable for the immediate initiation of the most adequate treatment for the patient.
Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016
Programa Estatal de I+D+i Orientada a los Retos de la Sociedad
Retos de Investigación: Proyectos de I+D+i
Gobierno De España. Ministerio De Economía Y Competitividad, Mineco
Ruiz, M.; Mujica, L.E.; Alferez, E.; Acho, L.; Tutivén, C.; Vidal, Y.; Rodellar, J.; Pozo, F. Mechanical systems and signal processing Vol. 107, p. 149-167 DOI: 10.1016/j.ymssp.2017.12.035 Date of publication: 2018-07 Journal article
Merino, A.; Puigví, L.; Boldú, L.; Alferez, E.; Rodellar, J. International journal of laboratory hematology Vol. 40, num. S1, p. 54-61 DOI: 10.1111/ijlh.12832 Date of publication: 2018-05-09 Journal article