Abstract
Background:
Therapy-related myeloid neoplasms (t-MN) are a leading cause of non-relapse mortality after autologous peripheral blood stem cell transplantation (aPBSCT) in patients with Hodgkin lymphoma (HL) and non-Hodgkin lymphomas (NHL). t-MN patients treated at an earlier stage of disease evolution have a better prognosis, presenting a need to identify patients at risk for t-MN.
Methods:
Using a prospective longitudinal study design, we evaluated peripheral blood parameters pre-aPBSCT and on Day 100, 6mo, 1y, 2y and 3y in 304 patients treated with aPBSCT. We examined the relation between peripheral blood parameters and subsequent development of t-MN. We developed nomograms to identify patients at risk for t-MN.
Results:
Twenty-one patients developed t-MN at a median of 1.95y post-aPBSCT. Hemoglobin, hematocrit, white blood cell and platelet count were lower among patients who developed t-MN compared to those who did not; these differences appeared soon after aPBSCT, persisted and preceded development of t-MN. Older age at aPBSCT (HRper_year_increase=1.08, p=0.007), exposure to TBI (HR=2.90, p=0.04) and low 100d platelet count (HRincrease_per_unit_decline_in_PLT=1.01, p=0.002) predicted subsequent t-MN. These parameters and primary diagnosis allowed identification of patients at high risk of t-MN (e.g., HL patient undergoing aPBSCT at age 70 with TBI, and with a Day 100 PLT between 100k and 150k, would have a 62% probability of developing t-MN at 6y post-aPBSCT).
Conclusions:
Abnormalities in peripheral blood parameters can identify patients at high risk for t-MN after aPBSCT for HL or NHL, allowing opportunities to personalize close surveillance and possible disease-modifying interventions.
Keywords: Therapy-related myeloid neoplasms, autologous peripheral blood stem cell transplantation, lymphoma, peripheral blood parameters
Précis
In this longitudinal study of lymphoma patients treated with autologous peripheral blood stem cell transplantation, blood parameters were altered among patients who subsequently developed therapy-related leukemia compared to those who did not. These differences appeared soon after transplantation, persisted and preceded development of t-MN, allowing identification of those at high risk of therapy-related leukemia and providing opportunities to personalize close surveillance and possible disease-modifying interventions among those at highest risk.
INTRODUCTION
Therapy-related myeloid neoplasms (t-MN) are a major cause of non-relapse mortality after autologous peripheral blood stem cell transplantation (aPBSCT) for Hodgkin lymphoma (HL) or non-Hodgkin lymphoma (NHL).1 The median time to development of t-MN is 12mo to 24mo after aPBSCT (range 4mo to 6y). The prognosis of t-MN treated with conventional therapy is uniformly poor, with a median survival of 6mo; patients salvaged with allogeneic blood or marrow transplantation (BMT) experience survival rates up to 24% at 3y.2–5 Patients with advanced disease have a poor prognosis.2, 5, 6
t-MN represent a cumulative consequence of exposure to alkylating agents and/or topoisomerase II inhibitors used to treat HL/NHL7–10 or to mobilize peripheral blood stem cell autografts7, 9, aggravated by proliferative stress in the immediate post-aPBSCT period11. t-MN are a clonal disorder affecting an early hematopoietic progenitor cell.12 Clonal hematopoiesis (defined as presence of myeloid malignancy-associated clonal mutations) in the PBSC product is associated with increased risk of t-MN.13 Additionally, mutated hematopoietic stem cell clones resist chemotherapy and expand preferentially after treatment.14
Although dysplastic changes in the marrow after aPBSCT are informative, especially if associated with cytogenetic abnormalities, development of dysplasia and cytogenetic abnormalities occur late in the evolutionary spectrum, which may preclude timely therapeutic interventions. Transplant clinicians often observe cytopenias after aPBSCT, but presence of cytopenias alone is not sufficient for a diagnosis of t-MN. In non-transplant settings, individuals with unexplained cytopenias are at a modest risk of progression to myelodysplastic syndrome (MDS) or acute myelogenous leukemia (AML); the risk is higher in patients who have accompanying mutations (clonal cytopenia of undetermined significance [CCUS]) than in patients without mutations (idiopathic cytopenia of undetermined significance [ICUS]).15 There is no consensus regarding selection of patients for assessment of clonal hematopoiesis and risk for t-MN, and performing sequencing in all is logistically impractical.
In this report, we investigated whether alterations in different components of the complete blood counts (CBC) precede development of t-MN after aPBSCT. We aimed to describe longitudinal trends in specific peripheral blood parameters from pre-aPBSCT to a timepoint post-aPBSCT, but prior to the development of t-MN, and to compare these trends with those in patients from the concurrent cohort who did not develop t-MN. Additionally, we examined baseline and Day 100 characteristics to identify patients at increased risk for developing t-MN.
MATERIALS AND METHODS
Patient Selection, Sample Collection, Data Abstraction
Using a prospective longitudinal study design, we conducted a serial evaluation of peripheral blood parameters pre-aPBSCT and at pre-determined timepoints after aPBSCT (100d, 6mo, 1y, 2y, and 3y). The peripheral blood parameters included red cell distribution width (RDW), hematocrit (HCT), hemoglobin (HGB), mean corpuscular volume (MCV), white blood cell count (WBC), and platelet count (PLT). We obtained these peripheral blood parameters for 304 consecutive patients who underwent aPBSCT for HL or NHL at a single transplant center. Twenty-one patients developed t-MN after aPBSCT (cases), while the remaining 283 patients served as controls. We excluded peripheral blood parameter values obtained within 3mo of relapse of HL/NHL, development of t-MN, solid subsequent malignant neoplasms (SMNs), or death. This study was approved by the institutional review board in accordance with an assurance filed with and approved by the Department of Health and Human Services, and met all requirements of the Declaration of Helsinki.
Using medical records, we abstracted age at aPBSCT, sex, race/ethnicity, pre-aPBSCT therapeutic exposures, conditioning, PBSC mobilization regimens, number of days to PBSC collections to achieve to ≥2×106 cells/kg CD34+ cell count, disease status after aPBSCT, age at diagnosis of t-MN, and age at last follow-up or death. For patients who developed t-MN, we reviewed pathology reports, including morphologic and cytogenetic characteristics, to verify the diagnosis in accordance with WHO 2016 Classification of Tumors of Hematopoietic and Lymphoid Tissues.16 Cases with t-MN were classified as those with chromosome 5 or 7 abnormalities (5/7 t-MN), those with 11q23 abnormalities (11q23 t-MN), and those with other abnormalities such as del(20q), del(13q), etc. (other t-MN).
Statistical Analyses
We used SAS version 9.4 (SAS Institute, Cary, NC) and RStudio (R version 4.1.1) to analyze data. All tests of statistical significance were two-sided. We used a linear mixed effects (LME) model with subject-level random effect to examine the longitudinal trends in peripheral blood parameters from before aPBSCT to 3y post-aPBSCT, censoring at 3mo prior to t-MN, HL/NHL relapse, solid SMN or death. Each post-aPBSCT timepoint was treated as a discrete variable. We used t-MN*timepoint interaction to determine whether the change in the blood parameters differed between those who subsequently developed t-MN and those who did not, from baseline to each post-aPBSCT time point. We adjusted the models for demographic and clinical variables that distributed differently between t-MN and non-t-MN patients.
We used Kaplan-Meier techniques to calculate the overall survival and event-free survival of the cohort. Death due to any cause was used as the event of interest for overall survival. For event-free survival, an event was defined as relapse of primary disease, development of t-MN or solid SMN, or death, whichever occurred first. We used multivariable Cox regression models for calculating relative risk estimates of the association between t-MN risk and peripheral blood parameters at time of aPBSCT and at Day 100. Potential variables examined in the multivariable model included age at aPBSCT, sex, race/ethnicity, primary cancer diagnosis (HL, NHL), number of days to collect ≥2.0×106/kg of CD34+ cells, exposure to total body irradiation (TBI), carmustine, cyclophosphamide and etoposide, and, peripheral blood parameter values (for the Day 100 model). Backward variable selection guided by rule of minimizing AIC (Akaike information criterion) was used to derive the final parsimonious model to predict the risk of t-MN. Internal validation was performed using the bootstrap method.
We used the ‘rms’ package in RStudio program to generate a nomogram for the final Cox regression predictive models.17 The use of nomogram is as follows. 1) The value corresponding to a specific subject is read on the scale for each variable. 2) The total points are calculated by adding up all the scores obtained in step 1). 3) The probability of an event at each endpoint corresponding to the total score of the subject is read on the probability scale. The probabilities of t-MN at 3y and 6y derived from the nomogram were verified and compared with the value of the probabilities estimated using the Cox regression model.
RESULTS
Clinical Characteristics
In Table 1, we summarize the cohort characteristics. The 304 patients in the cohort included 84 with HL and 220 with NHL. Median age at aPBSCT was 47y (range, 17–76) and the median length of follow-up of this cohort was 4.3y (0.1–11.1). Seventy-three percent of the cohort consisted of non-Hispanic whites, and 58.9% were male. Overall, 98.7% of the cohort had received etoposide for conditioning, 71.1% had received cyclophosphamide, 72.7% had received carmustine and 22.4% had received TBI. The median PBSC CD34+ cell dose was 5.7×106/kg. The overall survival at 9y was 69.7% (Supplemental Figure 1) and the event-free survival at 9y was 53% (Supplemental Figure 2). Twenty-one patients developed t-MN at a median of 1.95y from aPBSCT (0.5–9.5). As shown in Table 1, for patients who developed t-MN, the median age at aPBSCT was 57y (32–69). Forty-seven percent of these patients had received TBI for conditioning. Of the 21 patients with t-MN, 12 (57.1%) died after a median of 15.5 months from t-MN diagnosis (range: 1–42). The morphologic and cytogenetic characteristics of the 21 t-MN patients as well as the peripheral blood parameters at the time of diagnosis of t-MN are detailed in Supplemental Table 1.
Table 1.
Demographic and Clinical Characteristics of the Cohort
| Variable | Entire cohort (n=304) | No t-MN (n=283) | t-MN (n=21) |
|---|---|---|---|
| Age at transplantation (years) | |||
| Median (range) | 50.0 (18–77) | 49.0 (18–77) | 57.0 (32–69) |
| Time from transplantation to t-MN or date of last contact (for patients with no t-MN) (years) | |||
| Median (range) | 4.3 (0.1–11.1) | 4.5 (0.1–11.1) | 1.95 (0.5–9.5) |
| Primary diagnosis (n, %) | |||
| Hodgkin lymphoma | 84 (27.6%) | 80 (28.3%) | 4 (19%) |
| Non-Hodgkin lymphoma | 220 (72.4%) | 203 (71.7%) | 17 (81%) |
| Race/ethnicity (n, %) | |||
| Non-Hispanic Whites | 221 (72.7%) | 205 (72.4%) | 16 (76.2%) |
| Hispanics | 52 (17.1%) | 48 (17.0%) | 4 (19%) |
| Other | 31 (10.2%) | 30 (10.6%) | 1 (4.8%) |
| Sex (n, %) | |||
| Female | 125 (41.1%) | 118 (41.7%) | 7 (33.3%) |
| Male | 179 (58.9%) | 165 (58.3%) | 14 (66.7%) |
| Year of aPBSCT (n, %) | |||
| 1990–2004 | 118 (38.8%) | 105 (37.1%) | 13 (61.9%) |
| 2005–2009 | 186 (61.2%) | 178 (62.9%) | 8 (38.1%) |
| CD 34 counts (106cells/kg) | |||
| <6 | 161 (53.5%) | 151 (53.5%) | 10 (52.6%) |
| ≥6 | 140 (46.5%) | 131 (46.5%) | 9 (47.4%) |
| Days to ≥2×10 6cells/kg CD34 cell count | |||
| Median (Range) | 2 (1–15) | 2 (1–15) | 2 (1–11) |
| ≤2 | 193 (64.3%) | 182 (64.5%) | 11 (61.1%) |
| >2 | 107 (35.7%) | 100 (35.5%) | 7 (38.9%) |
| Conditioning agents (n, %) | |||
| Total body irradiation | 68 (22.4%) | 58 (20.5%) | 10 (47.6%) |
| Etoposide | 300 (98.7%) | 279 (98.6%) | 21 (100%) |
| Cyclophosphamide | 216 (71.1%) | 202 (71.4%) | 14 (66.7%) |
| Carmustine | 221 (72.7%) | 210 (74.2%) | 11 (52.4%) |
| Status | |||
| No event | 171 (56.3%) | 171 (60.4%) | |
| t-MN | 21 (6.9%) | 21 (100%) | |
| Relapse | 102 (33.6%) | 102 (36.0%) | |
| Death | 10 (3.3%) | 10 (3.5%) | |
aPBSCT denotes autologous peripheral blood stem cell transplantation; t-MN denotes therapy-related myeloid neoplasms
Peripheral Blood Parameters
As shown in Table 2 and Figure 1, the observed pre-aPBSCT values for red cell parameters, WBC and PLT were comparable between patients who subsequently developed t-MN (cases) and those who did not (controls). However, these values differed between cases and controls after aPBSCT. Differences appeared soon after aPBSCT, were persistent and preceded development of t-MN. We describe specific trends adjusted for age at aPBSCT, HL/NHL diagnosis, exposure to TBI and carmustine, and the number of days to collect ≥2.0×106/kg CD34 cells. Mean values for RDW were significantly higher for cases than for controls at Day 100 (17.32 vs. 15.13, p=0.03) after aPBSCT (Figure 1A). Mean HCT values were significantly lower for cases at Day 100 (31.42 vs. 35.46, p=0.0002), 6mo (33.41 vs. 36.99, p=0.0004), 1y (34.79 vs. 38.65, p=0.0008), 2y (35.14 vs. 39.29, p=0.003), and 3y (33.43 vs. 39.74, p=0.006) after aPBSCT (Figure 1B). Mean HGB levels were significantly lower for cases at Day 100 (11.03 vs. 12.24, p=0.0004), 6mo (11.37 vs. 12.75, p <0.0001), 1y (11.99 vs. 13.31, p=0.0003), 2y (12.29 vs. 13.52, p=0.009), 3y (11.57 vs. 13.48, 0.01) post-aPBSCT (Figure 1C). The adjusted LME analysis showed that MCV was significantly lower in the t-MN patients at the 1y post-aPBSCT timepoint (99.1 vs. 95.4, p=0.05) (Figure 1D). Mean WBC count was significantly lower in cases on Day 100 (3.58 vs. 4.45, p=0.007), and at 2y (4.23 vs. 5.83, p=0.02) post-aPBSCT (Figure 1E). Mean PLT counts were significantly lower in cases on Day 100 (100.53 vs. 169.61, p <0.0001), 6mo (113.53 vs. 179.82, p=0.0001), 1y (124.43 vs. 184.91, p=0.002), and 2y (141.88 vs. 197.37, p=0.02) (Figure 1F).
Table 2.
Peripheral blood parameters at the pre-defined time-points
| Mean (standard deviation) of blood parameters | ||||||
|---|---|---|---|---|---|---|
| RDW | HCT | HGB | MCV | WBC | PLT | |
| Pre-aPBSCT | ||||||
| Cases (n=20) | 16.75 (3.87) | 33.89 (4.66) | 11.92 (1.71) | 90.97 (6.01) | 5.54 (3.14) | 201.25 (118.7) |
| Controls (n=282) | 16.41 (3.08) | 34.12 (5.16) | 11.64 (1.77) | 89.24 (5.96) | 6.77 (5.08) | 227.91 (124.58) |
| p-value | 0.7 | 0.8 | 0.5 | 0.2 | 0.1 | 0.4 |
| Day 100 | ||||||
| Cases (n=19) | 17.32 (3.75) | 31.42 (4.21) | 11.03 (1.5) | 96.85 (4.68) | 3.58 (1.16) | 100.53 (58.6) |
| Controls (n=241) | 15.13 (2.13) | 35.36 (4.12) | 12.24 (1.41) | 95.88 (5.75) | 4.45 (2.09) | 169.61 (70.65) |
| p-value | 0.03 | 0.0002 | 0.0004 | 0.5 | 0.007 | <0.0001 |
| 6 months | ||||||
| Cases (n=17) | 16.96 (5.4) | 33.41 (4.25) | 11.37 (1.46) | 97.93 (6.26) | 4.06 (1.32) | 113.53 (58.08) |
| Controls (n=228) | 14.38 (1.58) | 36.99 (3.7) | 12.75 (1.31) | 95.24 (6.25) | 4.79 (1.94) | 179.82 (68.91) |
| p-value | 0.09 | 0.0004 | <0.0001 | 0.1 | 0.1 | 0.0001 |
| 12 months | ||||||
| Cases (n=14) | 14.55 (1.57) | 34.79 (3.46) | 11.99 (1.25) | 99.1 (8.54) | 4.51 (1.41) | 124.43 (50.75) |
| Controls (n=194) | 13.91 (1.18) | 38.65 (3.84) | 13.32 (1.32) | 95.4 (5.88) | 5.02 (1.46) | 184.91 (70.58) |
| p-value | 0.09 | 0.0008 | 0.0003 | 0.05 | 0.2 | 0.002 |
| 24 months | ||||||
| Cases (n=8) | 14.6 (1.46) | 35.14 (2.72) | 12.29 (1.13) | 95.87 (8.52) | 4.23 (1.25) | 141.88 (71.65) |
| Controls (n=171) | 13.76 (1.26) | 39.29 (3.59) | 13.52 (1.3) | 94.55 (5.7) | 5.83 (1.95) | 197.37 (66.17) |
| p-value | 0.09 | 0.003 | 0.009 | 0.6 | 0.02 | 0.02 |
| 36 months | ||||||
| Cases (n=3) | 14.53 (1.29) | 33.43 (2.95) | 11.57 (0.92) | 95.47 (10.52) | 4.7 (1.6) | 138 (26.91) |
| Controls (n=154) | 13.74 (1.29) | 39.4 (3.71) | 13.48 (1.32) | 95 (5.36) | 5.96 (2.1) | 199.43 (66.38) |
| p-value | 0.3 | 0.006 | 0.01 | 0.9 | 0.3 | 0.1 |
aPBSCT denotes autologous peripheral blood stem cell transplantation; RDW denotes red cell distribution width; HCT denotes hematocrit; HGB denotes hemoglobin; MCV denotes mean corpuscular volume; WBC denotes white blood cell count; PLT denotes platelet count
Figure 1.
Temporal trends in peripheral blood parameters by t-MN status. Error bars reflect standard deviation
Figure 1A: RDW; Figure 1B: hematocrit; Figure 1C: hemoglobin; Figure 1D: MCV; Figure 1E: WBC; Figure 1F: platelet count.
Identifying patients at risk for t-MN
Predictors of t-MN using variables at time of aPBSCT:
Predictors of t-MN included age at aPBSCT (HRper_year_increase=1.09, 95%CI, 1.03–1.14, p=0.001) and TBI (HR=4.98, 95%CI, 2.00–12.36, p=0.0005) (Table 3). Predictors of 5/7 t-MN included age at aPBSCT (HRper_year_increase=1.14, 95%CI, 1.05–1.24, p=0.001) and TBI (HR=7.72, 95%CI, 2.20–27.10, p=0.001) (Table 3). Of note, pre-aPBSCT blood parameter values were not associated with t-MN risk.
Table 3.
Predictors of t-MN after aPBSCT for Hodgkin lymphoma or non-Hodgkin lymphoma
| Variables | HR | 95% CI | p-value | HR | 95% CI | p-value |
|---|---|---|---|---|---|---|
| All t-MN types | t-MN with chromosome 5/7 abnormalities | |||||
| Including patient characteristics at aPBSCT | ||||||
| Age at aPBSCT | ||||||
| Per year increase in age | 1.09 | 1.03–1.14 | 0.001 | 1.14 | 1.05–1.24 | 0.001 |
| Diagnosis | ||||||
| Non-Hodgkin lymphoma | REF | REF | ||||
| Hodgkin lymphoma | 1.66 | 0.52–5.23 | 0.4 | 1.94 | 0.40–9.33 | 0.4 |
| Total body irradiation | ||||||
| No | REF | REF | ||||
| Yes | 4.98 | 2.00–12.36 | 0.0005 | 7.72 | 2.20–27.10 | 0.001 |
| Including blood parameter values from day 100 | ||||||
| Age at aPBSCT | ||||||
| Per year increase in age | 1.08 | 1.02–1.14 | 0.007 | 1.16 | 1.06–1.28 | 0.002 |
| Primary cancer diagnosis | ||||||
| Non-Hodgkin lymphoma | REF | REF | ||||
| Hodgkin lymphoma | 2.29 | 0.71–7.45 | 0.2 | 5.67 | 1.00–32.19 | 0.05 |
| Total body irradiation | ||||||
| No | REF | REF | ||||
| Yes | 2.90 | 1.04–8.14 | 0.04 | 2.87 | 0.70–11.69 | 0.1 |
| Platelet count at day 100 | ||||||
| Per unit increase | 0.99 | 0.98–1.00 | 0.002 | 0.98 | 0.97–0.99 | 0.004 |
aPBSCT denotes autologous peripheral blood stem cell transplantation; t-MN denotes therapy-related myeloid neoplasms; HR denotes hazard ratio
Predictors of t-MN using Day 100 variables:
Given significant differences between cases and controls in key peripheral blood parameters observed as early as Day 100, we examined the association between Day 100 blood parameters and subsequent t-MN. Notably, Day 100 bone marrow biopsy did not show signs of dysplasia or cytogenetic abnormalities. As shown in Table 3, we identified low platelets at 100d (HRincrease_per_unit_decline_in_PLT=1.01, p=0.002), older age at aPBSCT (HRper_year_increase=1.08, p=0.007) and TBI (HR=2.90, p=0.04) to be associated with increased risk of any t-MN. Statistically significant associations were identified between low platelets at 100d (HRincrease_per_unit_decline_in_PLT=1.02, p=0.004), older age at aPBSCT (HRper_year_increase=1.16, p=0.002), and a primary diagnosis of HL (HR=5.67, p=0.05) and 5/7 t-MN.
Using age at aPBSCT (y), primary diagnosis (HL/NHL) and receipt of TBI (yes/no), we developed a nomogram that predicts t-MN after aPBSCT for HL or NHL. The nomogram was characterized by one scale corresponding to each variable, a score scale, a total score scale and a probability scale (Figure 2A). Using the nomogram, the probability of developing t-MN in a patient with HL who underwent aPBSCT at age 60y and received TBI, would be 32% at 3y post-aPBSCT and 42% at 6y post-aPBSCT (Figure 2B). Next, we developed a nomogram using the baseline variables (age at aPBSCT [y], primary diagnosis [HL/NHL], receipt of TBI [yes/no]) and Day 100 PLT [<50k, 50–100k, 100–150k and >150k]) (Figure 3A). Using the nomogram, the probability of developing t-MN in a patient with HL who underwent aPBSCT at age 70y, received TBI and had a PLT of 125k would be 46% at 3y post-aPBSCT and 62% at 6y post-aPBSCT (Figure 3B).
Figure 2A.
Risk prediction nomogram for t-MN post-aPBSCT for HL or NHL, incorporating baseline factors: age at aPBSCT (y), primary diagnosis (HL or NHL), and exposure to TBI (no or yes).
Figure 2B.
Risk prediction nomogram in a patient with t-MN after autologous BMT for HL. A line is drawn downward from the value of each category to the score line. The points are then added to determine the total score and a line is drawn upward to find the risk of t-MN. T-MN probability estimation: age at aPBSCT: 60y – score=70; diagnosis: HL – score=10; TBI: yes – score=30. Total score=110, with a t-MN probability at 3y of 32% and at 6y of 42%.
Figure 3A.
Risk prediction nomogram for t-MN incorporating baseline factors and day 100 peripheral blood parameters: age at aPBSCT (y), primary diagnosis (HL or NHL), exposure to TBI (no or yes), and Day 100 PLT (<50; 50–100; 100–150; >150).
Figure 3B.
Risk prediction nomogram for t-MN using baseline and day 100 data. T-MN probability estimation: age at aPBSCT: 70y – score=86; diagnosis: HL – score=19; TBI: yes – score=26; PLT at D100: 100–150 – score=16. Total score=130, with a t-MN probability at 3y of 46% and at 6y of 62%.
DISCUSSION
aPBSCT remains a viable curative option for patients with HL or NHL who relapse or are at high risk of relapse. However, aPBSCT for HL or NHL may be complicated by t-MN – a clonal disorder affecting an early hematopoietic progenitor cell and a major cause of non-relapse mortality. Patients with overt t-MN have a dismal prognosis with median overall survival measurable in months.6 Current induction therapies for t-MN produce only short-lived remissions, and allogeneic transplantation is the only approach that allows long-term disease remission. Patients with advanced disease have poorer prognosis after allogeneic transplantation2, 4, and treatment at an earlier stage of disease evolution yields better outcomes. Identifying aPBSCT recipients at high risk of t-MN could guide targeted surveillance and detection, and allow timely action in patients at high risk to develop t-MN. We address this critical need by determining whether peripheral blood parameters could identify aPBSCT recipients at risk for t-MN. Routine CBC is inexpensive and is part of standard follow-up assessment after aPBSCT. In this study, we identify alterations in CBC parameters associated with increased risk of t-MN in HL/NHL patients treated with aPBSCT. It is important to note that the onset of peripheral blood parameters alterations occurs as early as Day 100 in patients who eventually developed t-MN, and these findings could identify patients at risk for t-MN.
While the peripheral blood parameters at baseline did not differ between those who did and did not subsequently develop t-MN, we observed significantly lower hematocrit, hemoglobin, white blood cell and platelet counts as early as 100 days post-aPBSCT among patients who subsequently developed t-MN. Further, the trajectory of these blood parameters remained significantly lower through the 3y of follow-up or until 3 months prior to t-MN, whichever occurred first. The observed anemia and thrombocytopenia likely represent ICUS or CCUS, defined as states without or with clonal hematopoiesis where cytopenia cannot be attributed to other disease.15 These entities have a propensity of progression to MDS.18 We are currently conducting deep sequencing at several time points to evaluate whether cytopenias in our study are clonal and dynamic.
We found that older age at aPBSCT and use of TBI were associated with an increased risk of t-MN; these findings confirm previous studies that identified TBI19 and older age at aPBSCT20 to be associated with t-MN risk. Patients with HL were at higher risk, although the association did not reach statistical significance. Using these parameters, we developed a nomogram to identify patients at the time of aPBSCT, who would be at high risk for developing t-MN. For example, when a patient with HL receives aPBSCT with TBI at age 60, the risk of developing t-MN at 3y is 32% and at 6y is 42%. If such a patient were not to receive TBI, then the probability of t-MN would be 7.5% at 3y and 10% at 6y. Indeed, this nomogram allows a clinician to determine the risk of t-MN based on the age of the patient and their primary cancer diagnosis (HL or NHL), and then determine the additional risk by conditioning with TBI.
The median time from aPBSCT to onset of t-MN was approximately 2y, but with a wide range (0.5y to 9.5y). Given that we observed changes in blood parameters as early as 100 days post-aPBSCT, we used the Day 100 blood parameters to identify those at highest risk for developing t-MN. Our results indicate that patients with HL, those who were older at aPBSCT, those who had received TBI, or had low Day 100 platelet count were at the highest risk of t-MN. Including these parameters into the nomogram resulted in identifying patients at high risk for t-MN among those who were 100 days post-aPBSCT. Indeed, in a hypothetical situation with a patient with extreme characteristics, a HL patient who underwent aPBSCT at age 70y, received TBI, and had a Day 100 PLT between 100 and 150, would be projected to have a 62% probability of developing t-MN at 6y post-aPBSCT.
Incorporation of additional biological and genomic parameters may further increase accuracy of risk prediction. We believe that identifying those at highest risk prior to aPBSCT or at Day 100 would allow time to institute surveillance strategies or additional appropriate interventions for those at risk for t-MN. Examples of possible interventions include treatment with hypomethylating agents and potentially curative consolidation with allogeneic transplantation.
A major strength of our study is the prospective longitudinal cohort design, allowing evaluation of parameters at multiple time-points, including prior to aPBSCT, and allowing evaluation of dynamic changes in blood count parameters and their association with t-MN. The relatively small number of events and focus on only lymphoma patients limits our study. Similar studies in patients undergoing aPBSCT for multiple myeloma, currently the most common indication for aPBSCT, would be of interest. Further, we did not have pre- or post-aPBSCT sequencing data. Correlation of these blood parameters with the sequencing data would allow for a clearer understanding of the mechanism of evolution of t-MN as well as a refinement of the risk prediction model.
In conclusion, we have described dynamic changes of peripheral blood parameters and their association with the subsequent development of t-MN after aPBSCT in patients with HL and NHL. We identified older age at aPBSCT, use of TBI-based conditioning and thrombocytopenia on Day 100 post-aPBSCT, to be associated with increased risk of developing t-MN. We were able to use these variables to develop nomograms at baseline and on Day 100 to identify patients at high risk for t-MN. We hope that findings from this study will help identify patients who could benefit from deep sequencing to detect the prevalence of somatic mutations to guide intensity of surveillance and providing opportunity for interventions.
Supplementary Material
Funding
This study was supported in part by grants from the National Institute of Health (P50 CA107399 [S.J.F.]), and the Leukemia Lymphoma Society (2192 [S.B.])
Footnotes
Disclosure of Conflicts of Interest
There were no conflicts of interest noted.
References
- 1.Bhatia S, Robison LL, Francisco L, et al. Late mortality in survivors of autologous hematopoietic-cell transplantation: report from the Bone Marrow Transplant Survivor Study. Blood. 2005;105: 4215–4222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Friedberg JW, Neuberg D, Stone RM, et al. Outcome in patients with myelodysplastic syndrome after autologous bone marrow transplantation for non-Hodgkin’s lymphoma. J Clin Oncol. 1999;17: 3128–3135. [DOI] [PubMed] [Google Scholar]
- 3.Witherspoon RP, Deeg HJ. Allogeneic bone marrow transplantation for secondary leukemia or myelodysplasia. Haematologica. 1999;84: 1085–1087. [PubMed] [Google Scholar]
- 4.Yakoub-Agha I, de La Salmoniere P, Ribaud P, et al. Allogeneic bone marrow transplantation for therapy-related myelodysplastic syndrome and acute myeloid leukemia: a long-term study of 70 patients-report of the French society of bone marrow transplantation. J Clin Oncol. 2000;18: 963–971. [DOI] [PubMed] [Google Scholar]
- 5.Anderson JE, Gooley TA, Schoch G, et al. Stem cell transplantation for secondary acute myeloid leukemia: evaluation of transplantation as initial therapy or following induction chemotherapy. Blood. 1997;89: 2578–2585. [PubMed] [Google Scholar]
- 6.Fianchi L, Pagano L, Piciocchi A, et al. Characteristics and outcome of therapy-related myeloid neoplasms: Report from the Italian network on secondary leukemias. Am J Hematol. 2015;90: E80–85. [DOI] [PubMed] [Google Scholar]
- 7.Krishnan A, Bhatia S, Slovak ML, et al. Predictors of therapy-related leukemia and myelodysplasia following autologous transplantation for lymphoma: an assessment of risk factors. Blood. 2000;95: 1588–1593. [PubMed] [Google Scholar]
- 8.Howe R, Micallef IN, Inwards DJ, et al. Secondary myelodysplastic syndrome and acute myelogenous leukemia are significant complications following autologous stem cell transplantation for lymphoma. Bone Marrow Transplant. 2003;32: 317–324. [DOI] [PubMed] [Google Scholar]
- 9.Kalaycio M, Rybicki L, Pohlman B, et al. Risk factors before autologous stem-cell transplantation for lymphoma predict for secondary myelodysplasia and acute myelogenous leukemia. J Clin Oncol. 2006;24: 3604–3610. [DOI] [PubMed] [Google Scholar]
- 10.Govindarajan R, Jagannath S, Flick JT, et al. Preceding standard therapy is the likely cause of MDS after autotransplants for multiple myeloma. Br J Haematol. 1996;95: 349–353. [DOI] [PubMed] [Google Scholar]
- 11.Bhatia R, Van Heijzen K, Palmer A, et al. Longitudinal assessment of hematopoietic abnormalities after autologous hematopoietic cell transplantation for lymphoma. J Clin Oncol. 2005;23: 6699–6711. [DOI] [PubMed] [Google Scholar]
- 12.Mach-Pascual S, Legare RD, Lu D, et al. Predictive value of clonality assays in patients with non-Hodgkin’s lymphoma undergoing autologous bone marrow transplant: a single institution study. Blood. 1998;91: 4496–4503. [PubMed] [Google Scholar]
- 13.Gibson CJ, Lindsley RC, Tchekmedyian V, et al. Clonal Hematopoiesis Associated With Adverse Outcomes After Autologous Stem-Cell Transplantation for Lymphoma. J Clin Oncol. 2017;35: 1598–1605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature. 2015;518: 552–555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Steensma DP. The Clinical Challenge of Idiopathic Cytopenias of Undetermined Significance (ICUS) and Clonal Cytopenias of Undetermined Significance (CCUS). Curr Hematol Malig Rep. 2019;14: 536–542. [DOI] [PubMed] [Google Scholar]
- 16.Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127: 2391–2405. [DOI] [PubMed] [Google Scholar]
- 17.Iasonos A, Schrag D, Raj GV, Panageas KS. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol. 2008;26: 1364–1370. [DOI] [PubMed] [Google Scholar]
- 18.Valent P, Bain BJ, Bennett JM, et al. Idiopathic cytopenia of undetermined significance (ICUS) and idiopathic dysplasia of uncertain significance (IDUS), and their distinction from low risk MDS. Leuk Res. 2012;36: 1–5. [DOI] [PubMed] [Google Scholar]
- 19.Metayer C, Curtis RE, Vose J, et al. Myelodysplastic syndrome and acute myeloid leukemia after autotransplantation for lymphoma: a multicenter case-control study. Blood. 2003;101: 2015–2023. [DOI] [PubMed] [Google Scholar]
- 20.Micallef IN, Lillington DM, Apostolidis J, et al. Therapy-related myelodysplasia and secondary acute myelogenous leukemia after high-dose therapy with autologous hematopoietic progenitor-cell support for lymphoid malignancies. J Clin Oncol. 2000;18: 947–955. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.