A process mining- deep learning approach to predict survival in a cohort of hospitalized COVID-19 patients.

BACKGROUND: Various machine learning and artificial intelligence methods have been used to predict outcomes of hospitalized COVID-19 patients. However, process mining has not yet been used for COVID-19 prediction. We developed a process mining/deep learning approach to predict mortality among COVID-19 patients and updated the prediction in 6-h intervals during the first 72 h after hospital admission. METHODS: The process mining/deep learning model produced temporal information related to the variables and incorporated demographic and clinical data to predict mortality. The mortality prediction was updated in 6-h intervals during the first 72 h after hospital admission. Moreover, the performance of the model was compared with published and self-developed traditional machine learning models that did not use time as a variable. The performance was compared using the Area Under the Receiver Operator Curve (AUROC), accuracy, sensitivity, and specificity. RESULTS: The proposed process mining/deep learning model outperformed the comparison models in almost all time intervals with a robust AUROC above 80% on a dataset that was imbalanced. CONCLUSIONS: Our proposed process mining/deep learning model performed significantly better than commonly used machine learning approaches that ignore time information. Thus, time information should be incorporated in models to predict outcomes more accurately.

Hemopoietic stem cell transplantation in childhood: reduction in mortality and improvement of survival over the years.

Stem cell transplantation (SCT) is an effective treatment for life-threatening hematologic and nonhematologic pediatric diseases. Reducing transplant-related mortality (TRM), a major complication of SCT, to improve long-term survival, therefore, is one of the main objectives of transplantation teams. We analyzed TRM and overall survival (OS) over the years in children undergoing SCT in our center. From June 1998 to October 2002, 156 consecutive children, 105 boys and 51 girls, median age 10 years (range, 2-18), with different diagnoses underwent SCT (100 autologous and 56 allogeneic). OS and TRM were analyzed in 2 different periods (June 1989-December 1998 and January 1999-October 2002) and grouped according to the different SCT modalities. The median follow-up was 18 months (range, 1-160). Autologous TRM showed a statistically significant improvement within 1999-2002 (0%) compared with 1989-1998 (12.2%) (P < .05). There were no statistical differences for allogeneic SCT. OS was 34% in the first period and 80.4% in the second period (P < .01), the improvement being for both autologous and allogeneic SCT. In our study, TRM decreased significantly for those children receiving autologous SCT in recent years. OS was significantly better in the latter period (1999-2002), both globally and for each SCT modality.

Bone marrow processing using the fenwal CS-3000 plus blood cell separator: results of 99 procedures.

BMT is used as an established therapy for patients with malignant and nonmalignant diseases. Many techniques for ex vivo treatment have been developed, but these techniques must be preceded by BM processing. We report our experience in processing 99 BM using the Fenwal CS-3000 Plus cell separator using the 1-special program. Ninety-nine procedures were performed in BM harvested from 73 patients and 26 healthy donors. The number of nucleated cells (NC), mononuclear cells (MNC), RBC, platelets, colony-forming units-granulocyte-macrophage (CFU-GM), CD34+ cells, relative purity of MNC and PMN, and volume were determined in the unprocessed BM and in the final product. BM processing resulted in NC, MNC, CFU-GM, and CD34+ cell recoveries of 31%, 82.2%, 117.6%, and 97.8%, respectively. RBC, PMN, platelets, and volume removal, respectively, were 96%, 92%, 37.2%, and 85.1%. In pediatric patients, the volume reduction was significantly lower than in adult patients (79.6% versus 88.8%). No other significant differences were found between pediatric and adult results. We conclude that BM processing with the Fenwal CS-3000 Plus cell separator provides a product that can undergo further ex vivo treatments or cryopreservation.

[Prospective comparative study of different amino acid and lipid solutions in parenteral nutrition of patients undergoing bone marrow transplantation]. / Estudio prospectivo comparativo de diferentes soluciones de aminoácidos y lípidos, en la nutrición parenteral de pacientes sometidos a trasplante de médula ósea.

UNLABELLED: The patient who will undergo a bone marrow transplant (BMT) has a high protein catabolism in the time period after the infusion of the marrow, and therefore there is a need for an adequate nutritional support. The objective of our study is to analyze the behavior of nutritional assessment parameters, the lipid metabolism, the number of days of mucositis, the number of infections, the number of days to recovery, and the number of hospitalization days when using different parenteral nutrition solutions: 22.5% and 45% branched chain amino acid solutions, and lipid solutions with long chain triglycerides (LCT), vs. medium chain triglycerides (MCT). MATERIAL AND METHODS: We have carried out a prospective, randomized study in patients who underwent a BMT who received parenteral nutrition. The supply of nitrogen was 1.5 +/- 0.3 g of AA/kg/day (either in standard solution or in a 45% branched chain AA solution). The caloric supply was similar in all the groups, with a proportion of 60% coming from carbohydrates and 40% from lipids, either LCT or MCT/LCT. The nutritional assessment parameters were studied, as well as of the lipid metabolism, and also clinical evolutive data: number of days of mucositis, number of days of PN, number of days hospitalized, number of infections, rate of infection density. All the data were measured and/or quantified 4 times: pretransplant, on day--of the transplant, and after 7 and 14 days after the transplant. RESULTS: 62 patients were studied. Group A: 19 patients treated with 22.5% branched chain amino acids + 20% LCT. Group B: 26 patients (45% branched chain amino acids + 20% LCT). Group C: 17 patients, (45% branched chain amino acids + 20% MCT/LCT). There is a quicker recover of the marrow in groups B + C: 14.4 vs. 11.7 and 11.1, with a p < 0.05. The nitrogen balance improves significantly in groups B and C (p < 0.05). The retinol-binding protein increases significantly from day 0 to day 7 (p < 0.01) in the LCT group (Group B). The phospholipids decrease in group B after one day (p < 0.05), and after the 7th day (p < 0.05). The triglycerides increase in group C between 7 and 14 days. The LDL/HDL quotient increases in group B after 14 days (p < 0.05). The triglycerides increase in group C between 7 and 14 days. The LDL/HDL quotient will increase in the B group after 14 days (p < 0.05). There are no differences in the number of days of mucositis, the total number of infections, the number of infections per 100 days of hospitalization, or in the number of hospitalization days. CONCLUSIONS: In patients who are given parenteral nutrition in the period immediately after the BMT, we found an improvement in the catabolic metabolism parameters when using a solution with a high proportion of branched chain amino acids (45%) and a smaller alteration of the metabolism of the plasmatic lipoproteins when we use MCT/LCT enriched solutions.

Philadelphia positive acute lymphoblastic leukemia 16 years after the apparent cure of acute lymphoblastic leukemia. New leukemia or late relapse?

A Philadelphia-positive ALL in an adult occurring 21 years after the initial diagnosis is reported here. This case raises the question as to whether or not this event is a relapse or a new leukemia. A possible role of interferon-alpha previously administered to the patient for a chronic viral hepatitis is discussed too.

Chronic lymphocytic leukemia with two cellular populations: a biphenotypic or biclonal disease.

A case of CLL with two different cellular populations is reported. A 50-year-old man was evaluated for persistent absolute lymphocytosis. A peripheral blood smear revealed numerous small lymphocytes (83% of white blood cells counted). Frequent Grumpecht shadows were present, too. On bone marrow aspirate smears lymphocytes comprised 85% of the total cells counted, and the bone marrow biopsy showed a mixed nodular-interstitial infiltration pattern. The immunophenotypic study showed two different leukemic populations. The first one (comprising 79% leukemic cells) was CD5+, CD19+, CD10-, CD20+, CD18-, CD22-, CD23+ +, lambda dim, and FMC7-. The second population (comprising 21% leukemic cells) was CD5+, CD19+, CD10-, CD20+, CD18+, CD22+, CD23+, lambda+ +, and FMC7+. Gene rearrangement studies detected the germline and one rearranged band in Jk blot with each restriction endonuclease. In the Jh blot the germline and two rearranged bands were detected with EcoRI and BamHI and three rearranged bands with HindIII. The JBI/JBII blot detected only the germline band. The detection of three rearranged bands was interpreted as evidence of the presence of at least two monoclonal populations of cells with the same light chain restriction.

Repopulation of circulating T, B and NK lymphocytes following bone marrow and blood stem cell transplantation.

A variety of T, B and natural killer (NK) cell subsets defined by surface markers were analyzed by double immunofluorescence flow cytometry in the peripheral blood of patients following autologous bone marrow transplantation (ABMT, n = 14), autologous peripheral blood stem cell transplantation (PBSCT, n = 10) and allogeneic bone marrow transplantation (allo-BMT, n = 6). Patients following ABMT were divided in 2 groups, those who did not received G-CSF post-transplant (ABMT, n = 6) and those who did (ABMT + G, n = 8). All patients following PBSCT or allo BMT received G-CSF. In all the groups prolonged significant decreases with respect to normal numbers were observed for the T CD3+, CD2+ and CD25+ subsets, more profound for the CD4+ subset but less for the CD8+ subset, especially following PBSCT (only decreased at 1 month). A significant expansion of the CD3+CD57+ and CD8+CD57+ phenotypes was noticed between 9 and 12 months following ABMT, the group of longer follow-up. Long-lasting expansion of the NK-like CD3+CD56+ and CD3+CD16+ subsets was also observed. The B CD19+ and CD20+ subsets had a significant overexpression from 4 months after ABMT, showing a normally balanced Igk+:Ig1+ ratio. Concordantly, the HLA-DR+ and HLA-DQ+ subsets showed significant increases. The NK CD56+ and CD16+ subsets had a faster recovery than the T or B subsets in all the groups. However, the CD3-CD56+, CD3-CD16+, CD16+CD56+, CD3-CD8+, and especially the CD3-CD57+, CD16+CD57+, and CD56+CD57+ subsets had a slower recovery than the global CD56+, CD16+, or CD57+ subsets. The biological and clinical implications of these findings are discussed.

[AB-cis phenotype with expression similar to AB-el]. / Fenotipo AB-cis con expresión similar a AB-el.

We studied a discrepancy between red blood cells and serum test. The red cell test was identified as an A group, and the serum test as an AB group. Then we performed adsorption-elution and saliva test, in order to demonstrate A, B and H substances. We found that the blood group was AB, and the B antigen belonged to a weak group named Bel. We studied her parent blood groups. Her mother was AB and her father 0. We conclude that the blood group belonged to an cis-AB phenotype and the B antigen was a weak Bel variant.

[Richter syndrome: morphologic transformation with persistence of the same phenotype]. / Síndrome de Richter: transformación morfológica con persistencia del mismo fenotipo.

Richter's syndrome is termed as the occurrence of a high-grade lymphoma along with a chronic lymphocytic leukaemia. We report a patient diagnosed as having a CLL evolving into an immunoblastic lymphoma. In spite of this morphologic change, the same identical immunophenotype remained in both types of cells assessed, that is, CLL-type lymphocytes and immunoblastic and lymphoplasmocytoid cells. This event appears to favour the common clonal origin of these two entities.

Lymphokine-activated killer cytotoxicity and lymphocyte subpopulations in patients with acute leukemia.

In this report we investigated lymphokine-activated killer (LAK) cytotoxicity and lymphocyte subpopulations of peripheral blood mononuclear cells (PBMCs) of patients with acute leukemia in complete remission after chemotherapy or autologous bone marrow transplantation (ABMT) and of normal donors. A positive linear correlation was found between the percentage of spontaneous LAK activity and that of lymphocyte subpopulations with phenotypes CD56+, CD3-CD8+ CD3-CD56+ and CD3-CD57+ in both groups of patients. A 3-day culture with IL-2 produced an up-regulation in the expression of the CD25, CD69, and HLA-DR markers proportional to the LAK cytotoxicity levels generated in the culture. Determination of percentages of spontaneous and in vitro generated LAK activity as well as of the above mentioned phenotypic markers contribute to the analysis of the process of LAK activation in patients with acute leukemia and may also be useful in those cases in which immunotherapy with IL-2 and/or LAK cells is anticipated.

[Immunotherapy with interleukin-2 and allogenic LAK cells in a patient with acute myeloblastic leukemia]. / Inmunoterapia con interleucina-2 y células LAK alogénicas en un paciente con leucemia aguda mieloblástica.

At the present time the effectiveness of interleukin-2 (IL-2) administered together with IL-2 in vitro activated lymphocytes (LAK) in the treatment of malignant neoplasias is being assessed. We report the case of an 8 -year- old patient suffering from acute myeloid leukemia (AML) who had relapsed from an autologous bone marrow transplant (ABMT) and received a second ABMT in partial remission. After the second ABMT, the patient, who entered into a phase of complete remission (CR) but with a severe thrombocytopenia, was treated with IL-2 and allogeneic LAK cells obtained from his father. During treatment the lymphocyte subpopulations and cytolytic activity were assayed. Increases in the patient's lymphocytes with CD3- CD4- CD8- CD16+, CD3- CD4- CD8- CD56+, CD3- CD4- CD8+ CD56+ and CD3- CD4- CD8+ CD56+ phenotypes were observed and these were correlated with cytolytic activity measured against the K562 cell line. The patient remained in CR for longer (14 months) than after the first ABMT (5 months). The immune activation produced by this therapy could have had an antileukemic effect in the patient.

Natural killer cytotoxicity and lymphocyte subpopulations in patients with acute leukemia.

We have studied lymphocyte subpopulations and the NK cytotoxicity of the PBMC from acute leukemia patients in complete remission after chemotherapy or ABMT and from normal donors. We have found a positive linear correlation between the percentage of subpopulations with CD3-CD16+, CD3-CD56+ and CD3-CD8+ phenotypes and the percentage of NK cytotoxicity. The slopes of the regression lines in the two groups of patients were lower than in normal donors, indicating a decreased ability of these cells to operate. The percentages of these subpopulations represent parameters for estimating the percentage of NK cytotoxicity both in normal donors and in patients with acute leukemia.

Generation of LAK cells in vitro in patients with acute leukemia.

The in vitro stimulation of lymphocytes with interleukin-2 (IL-2) generates lymphokine-activated killer (LAK) cells with tumoricidal potential. In this work we studied the cytolytic capacity of LAK cells in 51 acute leukemia patients in complete remission (CR) after chemotherapy (CT), in 24 acute leukemia patients who had undergone autologous bone marrow transplantation (ABMT), and in a control group of 44 normal donors. In the normal donor control group the effect of non-IL-2-activated peripheral blood mononuclear cells (PBMC) against blast cells was always lower than 10% lysis, which we have taken as a lower limit for positive results. In 95% of post-CT patients, the lytic effect of PBMC was negative. LAK cells produced positive results in 82% of normal donors and in 37.5% of post-CT patients. The effect of PBMC against K562, i.e. natural killer (NK) activity, in post-CT patients as well as in post-ABMT patients was reduced in comparison with the average for normal donors. LAK cells from 25% of post-CT patients had no notable activity against K562 or Raji, nor was there any positive effect against autologous blast cells. In the rest (75%), one-half generated positive activity. We did not observe any correlation between lytic activity in PBMCs or in LAK cells, nor did we observe significant differences between lytic activity in patients with acute lymphoblastic leukemia (ALL) and those with acute myeloblastic leukemia (AML), or between patients who had undergone CT and those receiving ABMTs. These results support the use of IL-2 as a treatment against minimal residual leukemia.

[Sudan black B-positive acute lymphoblastic leukemia]. / Leucemia aguda linfoblástica negro sudan B positivo.

Inasmuch as Sudan black B stain is highly specific for myeloid cells and since over 3% positive blast cells meet the needs of diagnosis for myeloblastic leukaemias, we had the opportunity to study a 27 year-old male with acute leukaemia whose morphological, cytochemical, immunocytochemical (HLA-DR, CD-10, CD-19, CD-20, CD-21 all positive) and ultrastructural features were clearly in accordance with acute lymphoblastic leukaemia, except for 42% Sudan black B positivity of the bone marrow blast cells. The practical interest of this case comes from the necessity to take this rare ALL case into account.

Post-remission therapy in acute myeloblastic leukemia. A randomized trial comparing early consolidation plus maintenance versus maintenance therapy alone.

During a 5-year period 203 previously untreated patients with acute myeloblastic leukemia entered an intensive induction chemotherapy regimen with daunorubicin, cytosine arabinoside, 6-thioguanine, vincristine and prednisone (DATOP). The complete remission rate was 64%. Patients in complete remission were randomly assigned to 3 courses of early consolidation with DATOP at lower dosage followed by maintenance chemotherapy, or to the same maintenance regimen in the absence of any consolidation courses. No significant differences were found between these options concerning disease-free survival (median 7.0 vs. 9.8 months; p greater than 0.10) or survival (median 15.8 vs. 19.4 months; p greater than 0.10). This study, in addition to the few previously reported randomized trials, suggests that early low-dose consolidation adds no benefit to maintenance chemotherapy in acute myeloblastic leukemia once complete remission has been achieved.