Decision-tree algorithm for optimized hematopoietic progenitor cell-based predictions in peripheral blood stem cell mobilization

Evaluation of blood cell count using an automatic hematology analyzer to optimize collection of peripheral blood progenitor cells by leukapheresis

BACKGROUND

Autologous stem cell transplantation is a treatment modality for several diseases. Prediction of successful mobilization may be useful to optimize hematopoietic stem cell collection.

STUDY DESIGN AND METHODS

This was a retrospective study with data from transplantation candidates between September 2015 and December 2021 being analyzed. The medical record of each patient was reviewed to mine mobilization information. The laboratory data analyzed were CD34+ cell enumeration and pre-collection peripheral blood cell count. The primary outcome, good mobilization, was defined as a CD34+ cell count ≥20/μL.

RESULTS

This study included 807 patients. Increased patient weight, low mean corpuscular volume, high nucleated red blood cells, peripheral blood mononuclear cell and immature granulocyte counts were significantly associated with good mobilization. In addition, patients diagnosed with multiple myeloma were two times more likely to be good mobilizers than patients with lymphoma. The model was applied to a validation set to identify patients who underwent apheresis (CD34+ cell count ≥10 µL), resulting in a sensitivity of 69 %, a specificity of 95 %, positive predictive value of 98 %, and a negative predictive value of 50 %.

CONCLUSION

Success in mobilization was greater in patients who underwent the first mobilization cycle and who had a diagnosis of multiple myeloma. Furthermore, higher body weight, and nucleated red blood cells, immature granulocytes and mononuclear cell counts, as well as low mean corpuscular volumes, were associated with successful mobilization.

Peripheral blood stem cell harvesting in young children weighing less than 15 kg

Autologous peripheral blood stem cell (PBSC) transplantation is crucial in pediatric cancer treatment, and tandem transplantation is beneficial in certain malignancies. Collecting PBSCs in small children with low body weight is challenging. We retrospectively analyzed data of pediatric cancer patients weighing <15 kg who underwent autologous PBSC harvesting in our hospital. Collections were performed in the pediatric intensive care unit over 2 or 3 consecutive days, to harvest sufficient stem cells (goal ≥2 × 106 CD34+ cells/kg per apheresate). From April 2006 to August 2021, we performed 129 collections after 50 mobilizations in 40 patients, with a median age of 1.9 (range, 0.6-5.6) years and a body weight of 11.0 (range, 6.6-14.7) kg. The median CD34+ cells in each apheresate were 4.2 (range, 0.01-40.13) × 106/kg. 78% and 56% of mobilizations achieved sufficient cell dose for single or tandem transplantation, respectively, without additional aliquoting. The preapheresis hematopoietic progenitor cell (HPC) count was highly correlated with the CD34+ cell yield in the apheresate (r = 0.555, P < 0.001). Granulocyte colony-stimulating factor alone was not effective for mobilization in children ≥2 years of age, even without radiation exposure. By combining the preapheresis HPC count ≥20/μL and the 3 significant host factors, including age <2 years, no radiation exposure and use of chemotherapy, the prediction rate of goal achievement was increased (area under the curve 0.787).

Risk factors of early disease progression and decreased survival for multiple myeloma patients after upfront autologous stem cell transplantation

Multiple myeloma (MM) stands as the second most prevalent hematological malignancy, constituting approximately 10% of all hematological malignancies. Current guidelines recommend upfront autologous stem cell transplantation (ASCT) for transplant-eligible MM patients. This study seeks to delineate factors influencing post-ASCT outcomes in MM patients. Our cohort comprised 150 MM patients from Taipei Veterans General Hospital, with progression-free survival (PFS) as the primary endpoint and overall survival (OS) as the secondary endpoint. A Cox proportional hazards model was employed to discern potential predictive factors for survival. ASCT age ≥ 65 (hazard ratio [HR] 1.94, 95% confidence interval [CI] 1.08-3.47) and the presence of extramedullary disease (HR 2.53, 95% CI 1.53-4.19) negatively impacted PFS. Conversely, treatment response ≥ VGPR before ASCT (HR 0.52, 95% CI 0.31-0.87) and total CD34+ cells collected ≥ 4 × 106 cells/kg on the first stem cell harvesting (HR 0.52, 95% CI 0.32-0.87) were positively associated with PFS. For OS, patients with ISS stage III (HR 2.06, 95% CI 1.05-4.04), the presence of extramedullary disease (HR 3.92, 95% CI 2.03-7.58), light chain ratio ≥ 100 before ASCT (HR 7.08, 95% CI 1.45-34.59), post-ASCT cytomegalovirus infection (HR 9.43, 95% CI 3.09-28.84), and a lower conditioning melphalan dose (< 140 mg/m2; HR 2.75, 95% CI 1.23-6.17) experienced shorter OS. In contrast, post-ASCT day + 15 absolute monocyte counts (D15 AMC) > 500/µl (HR 0.36, 95% CI 0.17-0.79) and post-ASCT day + 15 platelet counts (D15 PLT) > 80,000/µl (HR 0.48, 95% CI 0.24-0.94) were correlated with improved OS. Significantly, early PLT and AMC recovery on day + 15 predicting longer OS represents a novel finding not previously reported. Other factors also align with previous studies. Our study provides real-world insights for post-ASCT outcome prediction beyond clinical trials.

Prediction of future gene expression profile by analyzing its past variation pattern

A number of initial Hematopoietic Stem Cells (HSC) are considered in a container that are able to divide into HSCs or differentiate into various types of descendant cells. In this paper, a method is designed to predict an approximate gene expression profile (GEP) for future descendant cells resulted from HSC division/differentiation. First, the GEP prediction problem is modeled into a multivariate time series prediction problem. A novel method called EHSCP (Extended Hematopoietic Stem Cell Prediction) is introduced which is an artificial neural machine to solve the problem. EHSCP accepts the initial sequence of measured GEPs as input and predicts GEPs of future descendant cells. This prediction can be performed for multiple stages of cell division/differentiation. EHSCP considers the GEP sequence as time series and computes correlation between input time series. Two novel artificial neural units called PLSTM (Parametric Long Short Term Memory) and MILSTM (Multi-Input LSTM) are designed. PLSTM makes EHSCP able to consider this correlation in output prediction. Since there exist thousands of time series in GEP prediction, a hierarchical encoder is proposed that computes this correlation using 101 MILSTMs. EHSCP is trained using 155 datasets and is evaluated on 39 test datasets. These evaluations show that EHSCP surpasses existing methods in terms of prediction accuracy and number of correctly-predicted division/differentiation stages. In these evaluations, number of correctly-predicted stages in EHSCP was 128 when as many as 8 initial stages were given.

Risk factors for poor mobilization in solid tumors: How effectively can we mobilize patients with solid tumors?

BACKGROUND

In the literature, risk factors for poor mobilization were tried to identify. However, most of the studies consisted heterogeneous group of patients including both hematologic and oncologic malignancies. In this study, we aimed to identify the risk factors for poor mobilization in adults with solid tumors.

METHODS

We enrolled 49(47 men, 2 women) adult patients with solid tumor who were mobilized between September 2007 and February 2017. All the mobilization procedures were performed with G-CSF(10μg/kg/day) with chemotherapy. Mobilization insufficiency was defined as peripheral blood CD34+stem cell number less than 10/μl and/or total collected CD34+cells less than 2.5×10 ⁶/kg.

RESULTS

The patients were divided into two groups, patients with successful mobilization at the first attempt(group 1, 36 patients,73.5%) and poor mobilizers (group 2, 13 patients 26.5%). Second and third mobilization attempt was needed in 11 and 2 patients, respectively. The median number of CD34+cells collected was 7,08×10⁶/kg(0,6-19) with a median 4(1-6) apheresis. There was no statistical difference between two groups in terms of patient's and mobilization characteristics. Only number of CD 34+stem cells collected was statistically different (median 9,07×10⁶/kg CD34+cells in group 1 versus 2,14×10⁶/kg in group 2, p<0.05). The only possible risk factor that we could define was presence of organ metastasis.

CONCLUSIONS

Since several methods and new drugs are available for peripheral stem cell collecting, risk factors should be identified clearly in adult population with solid tumors. So multicenter studies should be constructed for resolving this problem.

Identification of cell morphology parameters from automatic hematology analyzers to predict the peripheral blood CD34-positive cell count after mobilization

Optimal timing of apheresis initiation is important for maximizing the hematopoietic stem cell (HSC) yield. This study aimed to identify useful parameters from automatic hematology analyzers for predicting the peripheral blood CD34+ cell count after mobilization. We prospectively enrolled 53 healthy donors and 72 patients, and evaluated 43 cell morphology parameters from Unicel DxH800 (Beckman Coulter, USA) and Advia 2120i (Siemens Healthcare Diagnostics, USA). The correlation of each parameter with the CD34+ cell count in pre-apheresis blood samples was analyzed. The delta neutrophil index (DNI) from Advia 2120i, standard deviation of volume of neutrophils and monocytes (SD-V-NE and SD-V-MO), standard deviation of conductivity of neutrophils and monocytes (SD-C-NE and SD-C-MO), mean conductivity of neutrophils and monocytes (MN-C-NE and MN-C-MO), and standard deviation of axial light loss of neutrophils and monocytes (SD-AL2-NE and SD-AL2-MO) from DxH800 showed significant correlations with the CD34+ cell count. SD-V-NE, SD-C-NE, and SD-C-MO showed good or fair area under the curve values for the prediction of the CD34+ cell count. SD-V-NE, SD-C-NE, and SD-C-MO from DxH800 will provide rapid, useful information for the initiation of apheresis after mobilization.