scholarly journals Acute Myeloid Leukemia (AML) Detection Using AlexNet Model

Complexity ◽  
2021 ◽  
Vol 2021 ◽  
pp. 1-8
Author(s):  
Maneela Shaheen ◽  
Rafiullah Khan ◽  
R. R. Biswal ◽  
Mohib Ullah ◽  
Atif Khan ◽  
...  

Acute Myeloid Leukemia (AML) is a kind of fatal blood cancer with a high death rate caused by abnormal cells’ rapid growth in the human body. The usual method to detect AML is the manual microscopic examination of the blood sample, which is tedious and time-consuming and requires a skilled medical operator for accurate detection. In this work, we proposed an AlexNet-based classification model to detect Acute Myeloid Leukemia (AML) in microscopic blood images and compared its performance with LeNet-5-based model in Precision, Recall, Accuracy, and Quadratic Loss. The experiments are conducted on a dataset of four thousand blood smear samples. The results show that AlexNet was able to identify 88.9% of images correctly with 87.4% precision and 98.58% accuracy, whereas LeNet-5 correctly identified 85.3% of images with 83.6% precision and 96.25% accuracy.

Blood ◽  
1997 ◽  
Vol 89 (9) ◽  
pp. 3330-3334 ◽  
Author(s):  
Wa'el El-Rifai ◽  
Tapani Ruutu ◽  
Erkki Elonen ◽  
Liisa Volin ◽  
Sakari Knuutila

Abstract The presence of residual leukemic cells was studied using metaphase-fluorescence in situ hybridization (FISH) in 22 patients with acute myeloid leukemia treated with chemotherapy only or chemotherapy followed by allogeneic bone marrow transplantation. The patients were followed up during their complete remission (CR) for 4 to 108 months (median, 21 months). A total of 88 BM samples was studied. In most of the samples more than 1,000 metaphase cells were analyzed. Residual leukemic cells were detected in 9 of 22 patients (41%). All patients who had an increasing and/or persisting level of abnormal cells in two or more subsequent samples or whose initial samples contained more than 1% of abnormal cells relapsed with one exception, in whom the later subsequent samples showed disappearance of abnormal cells. The time span before the first positive sample seems to be insignificant with regard to the outcome of relapse. Absence or single occurrence of abnormal cells followed by their disappearance was in agreement with CR in all the cases (16 patients). Our results indicate that metaphase-FISH is a reliable tool in the quantitation of residual leukemic cells and provides valuable prognostic information for patients with AML.


Blood ◽  
2001 ◽  
Vol 97 (12) ◽  
pp. 3882-3889 ◽  
Author(s):  
Michaela Feuring-Buske ◽  
Donna E. Hogge

Efflux of Hoechst 33342 from normal hematopoietic cells identifies a “side population” (SP+) of negatively staining cells that, in the mouse, are largely CD34− and are enriched for primitive progenitors. To further characterize human SP+cells, blood or bone marrow from 16 patients with acute myeloid leukemia (AML) was analyzed for their presence, immunophenotype, and cytogenetic and functional properties, and for the relation between SP phenotype and multidrug resistance-1 (MDR-1) expression. The mean percentages of SP+ and MDR+ cells was 8.1% (range, 0.5%-29.9%) and 12.8% (range, 0%-54.8%), respectively, with no correlation between the 2 values. The percentages of SP+ cells that were CD34+CD38−, CD34+CD38+, or CD34− were 12% (range, 0.4%-50%), 25% (range, 0.5%-96%), and 63% (range, 4%-99%). Cytogenetically abnormal cells were always detected in the SP−CD34+CD38− and SP+CD34− fractions, and abnormal colonies (CFC), long-term culture-initiating cells (LTC-IC), and nonobese diabetic-severe combined immunodeficiency (NOD/SCID) mouse leukemia–IC were detected in the former fraction. No progenitors were detected among SP+CD34− cells in any of these assays from 9 of 10 samples. In contrast, exclusively normal cells were detected in the SP+CD34+CD38−fraction from 9 of 15 samples, and CFC, LTC-IC, and multilineage engraftment in NOD/SCID mice from this subpopulation were also cytogenetically normal in 6 of 8, 6 of 7, and 2 of 2 cases studied, respectively. In contrast to murine studies, primitive progenitors are enriched among SP+CD34+CD38− cells from patients with AML. The molecular basis for Hoechst dye efflux is uncertain because it does not appear to be related to MDR-1 expression.


Blood ◽  
2007 ◽  
Vol 110 (11) ◽  
pp. 2864-2864
Author(s):  
Apostolia M. Tsimberidou ◽  
Hagop M. Kantarjian ◽  
Stefan Faderl ◽  
Sherry Pierce ◽  
Emil J. Freireich ◽  
...  

Abstract Background: The early mortality rates in older patients (generally defined as those > 60 years) following standard therapy for acute myeloid leukemia (AML), has spurred numerous trials of “targeted” treatments, typically limited to this age group. Here we examine whether younger patients with Zubrod performance status (PS) 2 might also be candidates for these therapies. Patients and Methods: We reviewed the records of 1841 patients 40 years or older with newly diagnosed AML (no acute promyelocytic leukemia) who received cytarabine-containing therapy at M. D. Anderson from 1980 to 2007. We chose patients with the following pretreatment characteristics because they are conventional criteria for entry onto trials of new therapies: PS 0–2, bilirubin < 2 mg/dL, and creatinine < 2 mg/dL. We calculated death rates by age and PS at 28 and 42 days after initiation of induction therapy. Results: Induction death rates in patients with AML by age and performance status The 28-day death rates in patients age 50–59 years with PS 2 were similar to those in patients ≥ 60 years with PS 0–1, and the 42-day death rates in the patients age 50–59 years with PS 2 were similar to those in patients age 60–79 with PS 0–1. Conclusion: Because of the high death rates following AML therapy in patients age 50–59 with PS 2, these patients should be eligible for targeted treatments now limited to older patients. Age, yrs Performance Status No. of Patients Dead by day 28, % Dead by day 42 % 40–49 0–1 288 7 9 40–49 2 38 5 5 50–59 0–1 405 4 6 50–59 2 85 14 15 60–69 0–1 423 8 11 60–69 2 111 17 17 70–79 0–1 301 12 16 70–79 2 124 23 33 ≥80 0–1 45 13 20 ≥ 80 2 21 38 43


Sign in / Sign up

Export Citation Format

Share Document