Hematological Toxicity and Bone Marrow Dose-Volume Effects in Chemotherapy-Naïve Patients Treated With Whole Pelvic and Para-aortic Lymph Node Radiation Therapy For Pathologic Pelvic Lymph Node–Positive Prostate Cancer Authors: Julie den Hartog, Gert De Meerleer, Annouschka Laenen, Robin De Roover, Steven Joniau, Koen Slabbaert, Ann Janssens, Herlinde Dumez, Alexander Giesen, Kato Rans Published: 2026 ABSTRACT AbstractPurposeWhole pelvic radiation therapy combined with androgen-deprivation therapy improves prostate cancer-specific survival in pathologic pelvic lymph node–positive prostate cancer. Elective para-aortic radiation therapy may further improve clinical outcomes and is under investigation in the Para-Aortic Radiotherapy (PART) study (NCT03079323). This study investigated the effect of whole pelvic radiation therapy plus PART on hematological toxicity (HemT).Methods and MaterialsHemT and bone marrow (BM) dose-volume relationships were analyzed in the first 75 PART patients enrolled in the University Hospitals Leuven, all receiving 24 months of androgen-deprivation therapy or antiandrogen therapy. Hematological changes were assessed longitudinally. Associations were analyzed using linear models for repeated measures, with false discovery rate correction.ResultsAt 57-months median follow-up (median age, 66 years), no HemT caused treatment interruption. All blood parameters declined significantly after radiation therapy (all P < .01 at 1, 3, 6, and 24 months). Grade 4 toxicity was limited to lymphopenia (acute 7%, late 1%), grade 3 leukopenia occurred in 1% acutely and late. No grade 3-4 anemia or grade 2-4 thrombocytopenia was seen. BM dosimetric analysis showed that larger low-dose volumes (V3, V5, and V10) in the lower pelvis, whole pelvis, and whole pelvis plus retroperitoneal region were significant associated with greater late absolute lymphocyte count (ALC) decline. Every 100 cm3 increase in absolute volume correlated with a late ALC decrease of −0.04 to −0.08 × 109 /L (all false discovery rates; P ≤ .01).ConclusionsExtended-field radiation therapy is clinically well tolerated in chemotherapy-naïve pathologic pelvic lymph node–positive prostate cancer patients, with no toxicity-related treatment interruptions. However, persistent lymphopenia was observed. Dose-dependent association between low-dose BM irradiation (V3-V10) and late ALC decline supports further research into BM-sparing techniques to reduce long-term toxicity. Introduction Approximately 6% to 14% of prostate cancer patients present with regional lymph node involvement, which is associated with an increased risk of cancer-specific mortality.1,2 In patients with pathologically confirmed pelvic lymph node–positive disease (pN1), oncological outcomes are heterogeneous and are influenced by primary tumor characteristics such as pT stage and surgical margin status, as well as number of positive lymph nodes.3 Adding adjuvant radiation therapy (RT) to androgen-deprivation therapy (ADT) has been shown to improve survival,4, 5, 6, 7 but optimal management of pN1 disease remains under investigation. Given the likelihood that pN1 patients with a high burden of pelvic nodal metastases harbor retroperitoneal micrometastases,8 the Para-Aortic Radiotherapy (PART) study was initiated in Belgium in 2017. This study investigated whether adding para-aortic nodal irradiation up to the level of the renal vasculature to whole pelvic radiation therapy (WPRT) improves clinical relapse-free survival in patients with substantial pelvic nodal disease burden, as confirmed by lymph node dissection.9 Although expanding radiation fields aims to improve oncological outcomes, it also increases the dose to surrounding healthy tissues, potentially increasing the risk of treatment-related toxicity.10 The extension of WPRT with PART exposes nearly half of the active bone marrow (BM) to radiation, hereby making hematological toxicity (HemT) a key concern.11 Damage to the red BM, which harbors hematopoietic stem cells (HSCs), essential for producing mature blood cells, leads to reduced hematologic output and therefore causes HemT. Irradiation of circulating blood cells plays a role as well as (circulating) lymphocytes, which are highly radiosensitive cells, undergo radiation-induced apoptosis.12 In contrast, circulating platelets (PLT) are considered relatively radioresistant13 and red blood cells (RBCs), though may structurally be altered by radiation, are generally not substantially affected.14,15 RT-induced anemia and thrombopenia, therefore, are rather caused by BM damage.16 BM is a radiosensitive “organ,” with an estimated alpha/beta value of 10 Gy,17 and behaves as a parallel organ composed of multiple subunits where both irradiated volume and dose contribute to HemT. The severity of HemT is further influenced by the irradiated field, as regions rich in active hematopoietic marrow are hypothesized to be more susceptible to radiation-induced damage.18 High-dose exposure to BM can lead to irreversible HSC depletion,19 whereas lower doses may allow for functional recovery. Ionizing radiation not only directly damages HSCs, but also disrupts the BM microenvironment, compromising the supportive stromal niche, exacerbating HSC injury, and ultimately impairing hematopoietic tissue repair.20 HemT recovery also depends on other adaptive responses, such as increasing hematopoietic activity in nonirradiated BM regions and the migration of circulating progenitors that support in-field regeneration and restoration of marrow function.19,21 BM dosimetry and HemT after WPRT and extended-field irradiation have been extensively researched in cervical cancer patients.22 To our knowledge, such prospective research is lacking in chemotherapy-naïve prostate cancer patients. Limited available studies primarily consider oncologic outcomes or gastrointestinal and genitourinary adverse events, without focus on HemT and BM dose-volume parameters.23,24 This subgroup analysis of the PART study evaluated HemT in patients with pN1 prostate cancer treated with WPRT plus PART and hormonal therapy, and specifically investigated the associated dose-volume parameters. Methods and Materials Patient population This is an interim analysis of the prospectively collected PART study data set: a multicenter, nonrandomized phase 2 study investigating the effect of elective para-aortic nodal irradiation in addition to prostate (bed) and WPRT in pN1 prostate cancer patients. Eligible patients had substantial pelvic lymph node involvement (≥2 positive lymph nodes, a nodal positivity ratio >7%, or extracapsular extension). The trial protocol was published in 2018.9 All patients received 24 months of hormonal therapy, consisting of either ADT or antiandrogen therapy (AAT, type bicalutamide). The correlation between BM toxicity and RT dose distribution was not a prespecified endpoint. The analysis was initiated based on the clinical observation that extended RT fields had an impact on blood cell count. In this analysis, the first 75 patients enrolled at the PART in the University Hospitals Leuven were included. One patient, treated with methotrexate for rheumatoid arthritis, developed grade 2 anemia and grade 4 thrombocytopenia and was excluded because of unrelated BM suppression, leaving 74 patients for analysis. RT planning and delivery Target volumes were delineated using planning computed tomography and magnetic resonance imaging. RT was delivered using volumetric modulated arc therapy. Dose prescription was as follows: 70 Gray (Gy) in 35 fractions to the postprostatectomy prostate bed, or 67.25 Gy in 25 fractions to the prostate in situ, along with a minimum dose of 45 Gy in 25 fractions to the pelvic and retroperitoneal lymph nodes. Details on the target volumes, dose constraints, and planning techniques have previously been described.9 No BM-specific dose constraints were applied. Laboratory analysis Blood samples were prospectively collected at baseline, end of RT and follow-up visits (month 1, 3, 6, 9, and 12, then every 6 months up to 5 years). Availability of the whole blood samples and timing can be found in Fig. E1. Measured parameters included RBC (×1012/L), hemoglobin (Hgb, g/dL), hematocrit levels (, %), white blood cell (WBC, ×109/L), PLT (×109/L) and relative and absolute lymphocyte counts (RLC, percentage and ALC, ×109/L). As reticulocyte data were incomplete (available in 70% of patients at end of RT, 30% at 3 months, and 27% at 6 months), they were excluded. HemT was assessed using delta, which is calculated as the follow-up value minus the baseline value. Negative delta values indicate a decrease in blood cell count, while positive values an increase. HemT recovery was arbitrarily defined as return to a level no more than 5% lower than the baseline level. Acute HemT was defined as toxicity after baseline up to 3 months post-RT, late HemT as new or persisting toxicity beyond 3 months. Grading was based on National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03.25 The lowest observed value was designated as the nadir. BM volume definitions BM volumes were delineated on the planning computed tomography using bony structures as a surrogate for the BM (based on Mell et al. 26 using bone window (-400–800 HU). Sub-volumes included: lumbosacral spine, iliac crests, lower pelvis, whole pelvis, retroperitoneal and whole pelvis plus retroperitoneal (ALL) (Fig. 1). The body volume receiving the ≥2 Gy was contoured as an investigational anatomic surrogate for circulating blood exposure. Dose-volume histograms (DVH) were calculated in absolute (cm3) and relative values (%) for V5-V50 (Gy). Graphic presentation of the bone sub-sites delineation. Red: lower pelvis; Orange: iliac crests ; Yellow: lumbosacral spine; Green: retroperitoneal. Not depicted: whole pelvis and body volume. Statistical analysis Linear models for repeated measures were used to analyze the time evolution of the delta levels, modeling time as a categorical variable. A heterogeneous autoregressive residual covariance matrix was modeled to account for the longitudinal data structure. Results are presented as mean estimates per time point with 95% confidence intervals (CIs). A P value <.05 indicates a significant difference with baseline. Similar models were used to analyze the association between baseline characteristics or dosimetric variables, and delta values of blood parameters. Clinical variables included age, diabetes mellitus, active smoking, hormonal therapy (ADT vs AAT), and the use of methylprednisolone during treatment. For dosimetric variables, we provide both raw and multiple testing-corrected P values, given the large number of parameters tested. Correction for multiple testing is based on the false discovery rate.27 Kaplan-Meier estimates were used for the recovery rate over time. A Cox proportional hazard model was used to study the prognostic effects of baseline characteristics for recovery. Results are presented as hazard ratios (HR) with 95% CIs. Logistic regression models were used to study the prognostic effects of baseline characteristics for grade ≥2 toxicity. Results are presented as odds ratios (ORs) with 95% CIs. All tests are performed at a 2-sided 5% significance level. Analyses have been performed using SAS software (version 9.4 of the SAS System for Windows). Results Patient and treatment characteristics This cohort comprised 74 evaluable patients; their baseline characteristics are summarized in Table 1. The median age was 66 years (range, 49-78). The majority of the patients (91%) received RT to the pelvic and retroperitoneal lymph nodes in combination with prostate (7%) or prostate bed RT (84%). Seven patients (9%) were treated to the pelvic and retroperitoneal lymph nodes only (salvage setting), as they have been treated with RT to the prostate (bed) in the past. In this postoperative setting, no persistent involved nodes were observed, therefore no nodal boosts were delivered. Median follow-up was 57 months (range, 21-90). All patients were treated with 24 months of hormonal therapy. Neither second-generation androgen receptor pathway inhibitors (ARPIs) nor Docetaxel were prescribed. During follow-up, 60 ADT-treated patients returned to noncastrate testosterone levels (≥50 ng/dL), after a median of 32 months from start of hormonal therapy. Of these, 40 patients reached normal testosterone levels (≥300ng/dL). Testosterone data were missing for 3 patients, the remaining 4 remained castrated after a median of 55 months. Seven patients (9%) received methylprednisolone for RT-induced nausea and/or vomiting. This was administered once daily in a tapering schedule; 4 received a maximum dose of 32 mg and 2 received 16 mg. Table 1Baseline patient characteristics and treatment descriptionTable 1 dummy alt textCharacteristicN = 74Median age (range) (y)66 (49-78)Mean age (y)65Treatment field(#, %) Prostate bed + WPRT/PART62 (84%) Prostate + WPRT/PART5 (7%) WPRT/PART only7 (9%)Diabetes mellitus(#, %)11 (15%)Active smoker* (#, %)5 (7%)Hormonal therapy(#, %) Androgen deprivation therapy67 (91%) Antiandrogen therapy7 (9%)Methylprednisolone†(#, %) 32 mg5 (7%) 16 mg2 (3%)Abbreviations: PART = para-aortic radiation therapy; WPRT = whole pelvic radiation therapy.⁎Active smoker at the start of radiation therapy.†Administration once daily in a tapering schedule for radiation therapy-induced nausea and/or vomiting Baseline patient characteristics and treatment description Abbreviations: PART = para-aortic radiation therapy; WPRT = whole pelvic radiation therapy. Active smoker at the start of radiation therapy. Administration once daily in a tapering schedule for radiation therapy-induced nausea and/or vomiting Hematologic evolution and clinical outcomes Baseline blood counts are presented in Table 2. Details concerning anemia (specified by Hgb), leukopenia (WBC), thrombopenia (PLT) and lymphopenia (ALC) toxicity scoring are presented in Table 3. HemT did not result in any treatment interruption or discontinuation. Lymphopenia was the only observed grade 4 toxicity. Five patients (7%) developed grade 4 acute lymphopenia, two recovered to baseline, two persisted at grade 1. The remaining patient evolved from acute grade 3 lymphopenia to grade 4 by month three, coinciding with antibiotic treatment for a urinary tract infection, and persisted at grade 3 during follow-up. One patient developed late grade 4 lymphopenia at month six without recovery after a follow-up of 36 months. Patients with higher grade lymphopenia at baseline showed an increased likelihood of grade ≥2 lymphopenia two years post-RT (OR, 2.25; 95% CI: 1.01-4.98, p < 0.05). No grade 3-4 anemia or grade 2-4 thrombocytopenia was observed. One patient developed grade 3 acute and late leukopenia, this patient had a baseline grade 2 leukopenia and recovered to the same level during follow-up. Table 2Normal laboratory range, baseline blood levels at inclusion, and recovery for hemoglobin (Hgb, g/dL), white blood cell (WBC, ×109/L), relative and absolute lymphocyte count (RLC and ALC, percentage and ×109 /L), and platelets (PLT, ×109)Table 2 dummy alt textNormal rangeStudy populationMean baselinevalue (range)Baseline < lowerlaboratory limitRecoveryat last FUMedian TTR(mo) (95% CI)Hgb14.00-18.0014.18 (10.60-17.70)42%73%9 (3-24)WBC4.00-10.006.71 (2.56-12.99)5%53%42 (24-NR)RLC20.00-50.0025.79 (3.38-50.40)23%65%30 (12-36)ALC1.20-3.601.65 (0.30-3.09)22%30%NR (54-NR)PLT150-450251 (95-634)5%72%6 (3-9)Abbreviations: FU = follow-up; NR = not reached; TTR = time to recovery. Normal laboratory range, baseline blood levels at inclusion, and recovery for hemoglobin (Hgb, g/dL), white blood cell (WBC, ×109/L), relative and absolute lymphocyte count (RLC and ALC, percentage and ×109 /L), and platelets (PLT, ×109) Abbreviations: FU = follow-up; NR = not reached; TTR = time to recovery. Table 3Incidence of acute and late toxicity, with grading and reference levels for hemoglobin (Hgb, g/dL), absolute lymphocyte count (ALC, ×109 /L), white blood cell (WBC, ×109/L) and platelets (PLT, ×109) according to the National Cancer Institute Common Terminology Criteria for Adverse Events Version 4 scoring systemTable 3 dummy alt textGrading HgbGrading ALCGradingRangePercentage of patients in study (%) (baseline/acute/late)GradingRangePercentage of patients in study (%) (baseline/acute/late)Grade 0Within normal range58 / 1 / 5Grade 0Within normal range78 / 1 / 9Grade 1< LNN-1042 / 96 / 86Grade 1< LNN-0.815 / 11 / 31Grade 2< 10-80 / 3 / 7Grade 2< 0.8-0.55 / 26 / 36Grade 3< 8-4.90 / 0 / 0Grade 3< 0.5-0.21 / 55 / 20Grade 4Life threatening0 / 0 / 0Grade 4< 0.20 / 7 / 1Grade 5Death0 / 0 / 0Grade 5Death0 / 0 /0Missing0 / 0 / 1Missing0 / 0 / 1Grading WBCGrading PLTGradingRangePercentage of patients in study (%) (baseline/acute/late)GradingRangePercentage of patients in study (%) (baseline/acute/late)Grade 0Within normal range95 / 49 / 47Grade 0Within normal range95 /88 / 84Grade 1< LNN-34 / 31 / 34Grade 1< LNN-755 / 12 / 16Grade 2< 3-21 / 19 / 16Grade 2< 75-500 / 0 / 0Grade 3< 2-10 / 1 / 1Grade 3< 50-250 / 0 / 0Grade 4< 10 / 0 / 0Grade 4< 250 / 0 / 0Grade 5Death0 / 0 / 0Grade 5Death0 / 0 / 0Missing0 / 0 / 1Missing0 / 0 / 0Abbreviations: ALC = absolute lymphocyte count; Hgb = hemoglobin; PLT = platelets; WBC = white blood cell. Incidence of acute and late toxicity, with grading and reference levels for hemoglobin (Hgb, g/dL), absolute lymphocyte count (ALC, ×109 /L), white blood cell (WBC, ×109/L) and platelets (PLT, ×109) according to the National Cancer Institute Common Terminology Criteria for Adverse Events Version 4 scoring system Abbreviations: ALC = absolute lymphocyte count; Hgb = hemoglobin; PLT = platelets; WBC = white blood cell. All blood parameters showed a significant decrease after RT, illustrated in Figure 2, Figure 3. Erythropoiesis remained significantly declined up to month 48 (all P < .05), with the largest mean Hgb decline observed 1 month after treatment, reaching −1.97 g/dL below baseline (95% CI, −2.23-−1.72). For WBC, ALC and PLT, values remained significantly decreased throughout the 5-year follow-up (all P < .02), with nadirs occurring at the end of RT. Despite the decline, WBC and PLT mean values remained within normal reference ranges throughout follow-up. In contrast, erythropoiesis and ALC mean values fell below the lower reference limit after treatment. Clinical variables impacting blood parameters are shown in Tables E1 and E2. Greater late WBC decline was predicted by higher age (−0.07 per year; 95% CI, −0.11-−0.03; P < .01), diabetes mellitus (−1.10; 95% CI, −1.88-−0.32; P < .01) and active smoking (−1.22; 95% CI, −2.32-−0.13; P = .03). In contrast, active smoking was associated with higher late RLC (8.00; 95% CI, 3.13-12.87; P < .01), despite smokers and nonsmokers having similar baseline values. Furthermore, the use of ADT (vs AAT) was associated with a greater late decline in RLC (−4.65; 95% CI, −9.00-−0.30; P = .04). Mean delta (change from baseline) of blood parameters over time, with 95% CIs: (a) hemoglobin, (b) white blood cell, (c) absolute lymphocyte count , and (d) platelets. Baseline is indicated by horizontal line at 0. Abbreviation: end RT: last day of radiation therapy. Mean values of blood parameters over time, with 95% CIs: (a) hemoglobin, (b) white blood cell, (c) absolute lymphocyte count, and (d) platelets. Horizontal lines represent upper and lower reference limits. Abbreviations: BL: baseline, eRT: last day of radiation therapy. No significant differences in erythropoiesis decline occurred between ADT versus AAT groups in either the early (−2.49; P = .91) or the late phase (2.40; P = .88). Patients receiving RT limited to the pelvic and para-aortic lymph nodes (vs also prostate[bed] RT) started off with a lower mean baseline of ALC (1.17 vs 1.62 × 109 /L) but showed less decline both in early (0.40; 95% CI, 0.06-0.74; P = .02) and late follow-up (0.38; 95% CI, 0.11-0.66; P < .01). PLT decline was more pronounced in patients with diabetes both in early (−46.50; 95% CI, −80.91-−12.09; P < .01) and late phase (−40.85; 95% CI, −64.71-−16.98; P < .01). Conversely, late PLT values were higher in active smokers (49.74; 95% CI, 17.21-82.27; P < .01) and in those receiving methylprednisolone during treatment (34.81; 95% CI, 2.75-66.87; P = .03). Median time to recovery differed among blood cell types (Table 2, Fig. E2). PLT recovered first (median, 6 months), followed by Hgb (9 months), RLC (30 months), and WBC (42 months). Median ALC recovery was not reached. Diabetes mellitus reduced WBC recovery probability (HR, 0.23; 95% CI, 0.06-0.98; P < .05). Patients receiving ADT had slower lymphocyte recovery than those on AAT, although it did not reach statistical significance (HR, 0.14; P = .05), the effect might be clinically relevant given the strong point estimate. No significant difference was detected in Hgb recovery between patients who received ADT, compared to AAT (HR, 1.09; P = 0.84). Methylprednisolone use during treatment was associated with improved PLT count recovery (HR, 2.60; 95% CI, 1.16-5.81; P = 0.02). Dosimetric correlates of hematologic parameters Dosimetric analysis demonstrated dose-volume effects specifically for late lymphocytes decline, no parameters remained significant for early decline or recovery after false discovery rates correction (Table E3 and E4). Late ALC and RLC declines were linked to low-dose irradiation of large BM volumes. For the combined whole pelvis and retroperitoneal BM structure (ALL), higher absolute volumes receiving low-dose radiation (V3-V10) were predictors of late ALC decline with a mean change of −0.04 per 100 cm³ increase (P < .01). Figure 4 provides a visual representation of these negative regression slopes for absolute volumes of the BM ALL structure. Correspondingly, increased percentage volumes for V3-V10 also significantly predicted lower late ALC, with mean change of −0.02 per 1% increase (P ≤ .01). Regression slopes with 95% CIs showing the relationship between absolute volumes (cm3) of the combined pelvic and retroperitoneal bone marrow structure (BM_ALL) receiving specific radiation doses V3, V5, and V10 and the mean absolute lymphocyte count delta (change from baseline) at the 6-month timepoint. Similar associations were observed for the lower pelvis where larger absolute and relative volumes receiving V3-10 were also linked to more ALC decline. To the whole pelvis substructure, every 100 cm3 increase in absolute volume of V3, V5, or V10 significantly predicted more late decline in both ALC (mean change, −0.07-−0.08; P < .01) and RLC (mean change, −0.74-−0.81; P = .02). Similarly, each 1% increase in whole pelvis relative V3, V5, and V10 volumes was significantly associated with ALC (mean change, −0.02; P <.01) and RLC (mean change, −0.23-−0.25; P = .02) decline. No false discovery rates-significant associations were observed for erythropoiesis or PLT. No significant associations were found between the substructure body volume receiving the ≥2 Gy, used as a surrogate for circulating blood exposure, and hematologic decline or recovery. Discussion This interim analysis explores HemT in chemo-naïve pN1 prostate cancer patients who have been treated with elective PART in addition to WPRT (NCT03079323).9 In our patient population, HemT never resulted in treatment interruption or discontinuation. HemT appeared across all major blood cell lines. The differing nadir timings and recovery capacity after irradiation, reflect the distinct radiation responses of each hematopoietic compartment, attributable to their differential degrees of radiosensitivity and proliferation capacity.28 Red cells showing delayed decline owing to a long lifespan and slower turnover. In contrast, WBC and PLT depleted more quickly due to their shorter lifespan and faster turnover. The prolonged lymphopenia observed is consistent with previous pelvic RT data in pelvis-irradiated patients.29 Extensive dose-volume analysis highlighted significant associations between low-dose marrow irradiation and persistent RLC and ALC decline, demonstrating that broad low-dose exposure impairs lymphocyte production. Older age, diabetes mellitus , and smoking were associated with larger absolute WBC declines in our cohort, possibly reflecting underlying baseline inflammation and reduced hematopoietic reserve. Smoking has been shown to induce oxidative stress and systemic inflammation,30 thereby exacerbating treatment-related toxicities.31 However, smokers exhibited higher PLT values in late follow-up, which may be related to chronic smoking-induced increases in PLT production.32 Concurrent methylprednisolone use was associated with more rapid PLT recovery, consistent with the known in vitro stimulatory effects of glucocorticoids on megakaryocyte proliferation and thrombopoiesis.33 In addition, baseline ALC grade independently predicted persistent grade ≥2 ALC toxicity at 2 years post-RT, emphasizing the importance of baseline hematologic status as a risk stratifier. All patients in the PART study were prescribed 24 months of hormonal therapy being ADT in the majority of cases and AAT when ADT was refused by the patients. ADT itself can increase the risk of HemT independently, as testosterone is known to affect the hematological system by inducing the release of erythropoietin, thereby increasing BM activity and incorporating iron into RBC.34,35 A cohort study has demonstrated that the use of ADT nearly triples the risk of developing new-onset anemia, with a higher risk associated with longer use and recovery after discontinuation.36 Although ADT usage is an established independent risk factor for anemia development during hormonal therapy, no statistically significant differences in Hgb decline or recovery was detected between patients treated with ADT versus AAT in this cohort. However, this comparison is limited by the small size of the AAT subgroup (n = 7). As ARPIs such as abiraterone and enzalutamide are increasingly reimbursed and consequently combined with RT, their hematologic safety profile becomes clinically relevant. Current data indicate that ARPIs significantly improve oncologic outcomes when added to ADT in high-risk nonmetastatic prostate cancer patients treated with local definitive therapy,37 with most additional adverse events being nonhematologic (eg, hypertension, rash, and fatigue), and if given in monotherapy, they are even associated with a decreased risk of anemia.38 Persistent lymphopenia after RT has been reported across tumor types. Our findings align with previous reports indicating a substantial risk of lymphopenia after pelvic RT in chemotherapy-naïve prostate cancer patients.39,40 Compared with the cohort reported by Pavarini et al,39 we observed higher rates of grade ≥2 lymphopenia at 2 years post-RT (26% vs 9.2%) and grade ≥3 lymphopenia (5% vs 0.8%). This increased ALC toxicity may reflect the larger RT fields used in the PART study, increasing the irradiated volume of BM, circulating lymphocytes, and lymphoid tissue. Factors related to RT, such as a larger treatment volume and a higher number of fractions, have been associated with an increased risk of severe lymphopenia for certain tumor sites.41 Although this analysis was not powered to evaluate oncologic outcomes, concerns about the clinical consequences of radiation-induced lymphopenia remain relevant. A meta-analysis of Damen et al42 showed that radiation-induced lymphopenia is associated with decreased overall survival in multiple solid tumors. Although this analysis did not include prostate cancer, it highlights the potential prognostic importance of sustained lymphocyte suppression. During follow-up, 11 patients in our cohort developed metastatic disease requiring metastasis-directed or hormonal therapy, including 2 patients who progressed while still receiving ADT. Notably, none of the patients with metastasis in our cohort required chemotherapy, a poly ADP-ribose polymerase (PARP) inhibitor or Lutetium-177 Prostate Specific Membrane Antigen therapy, which may carry additional marrow-suppressive risks. In other tumor types, such as recurrent cervical cancer, prior pelvic RT has been shown to increase chemotherapy-related BM suppression.43 Although most blood parameters recovered in this cohort, ALC recovery was strongly impaired. The relative short follow-up precludes definitive conclusions about absence of clinical impact. This is particularly relevant in a high-risk prostate cancer population with long-term survivorship and nonneglectable probability that patients will need (palliative) systemic therapies later on. Subclinical BM injury, though not yet been demonstrated in prostate cancer, could decrease tolerance to future systemic treatments. In literature, dosimetry is a frequently discussed strategy to prevent radiation induced lymphopenia, as there are no reimbursed pharmacological supportive agents in this setting yet. Currently, the BM is rarely included in the treatment planning optimization. This provides an opportunity for emerging strategies, including marrow-sparing intensity modulated radiation techniques or proton therapy, to play a role in reducing BM dose and so limiting lymphocyte toxicity and better preserve immune function. Several limitations must be acknowledged. This interim analysis is exploratory and underpowered for hematologic toxicity endpoints. Multiple comparisons increase the risk of false positives, although false discovery rates control was applied. Not all confounders could be fully accounted for, such as body mass index, unreported medications, or transient infections. Conclusion In this interim analysis of chemo-naïve pN1 prostate cancer patients treated with whole pelvic and elective retroperitoneal radiation therapy, hematologic toxicity did not interrupt treatment. A pronounced and sustained decline in absolute lymphocyte count was observed, with only 30% of patients recovered to baseline during follow-up. Low-dose irradiation (V3-V10) across large bone marrow volumes at pelvic and retroperitoneal region, was significantly associated with late lymphocyte decline. These findings demonstrate the relevance of bone marrow dose distribution, particularly low-dose “bath” effects, in shaping lymphocyte outcomes and support the further research of marrow-sparing strategies in future treatment planning. Disclosures The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Declaration of AI and AI-Assisted Technologies in the Writing Process During the completion of this work, the corresponding author used Google Gemini for language assistance to improve clarity and consistency; generative AI assistance was not used during the research or other writing processes. The text and images in this manuscript are entirely our own. After using this tool/service, the author reviewed and edited the content as needed and takes full responsibility for the content of the publication. Footnotes Sources of support: G.d.M. received a research grant for the Para-Aortic Radiotherapy (PART) study, by the organization “Kom op tegen kanker” (“Stand up to cancer”), ref #0010048. Granting bodies were not involved in the design of the study and collection, analysis, and interpretation of data, or in the writing of the manuscript. Research data are stored in an institutional repository and will be shared upon request to the corresponding author. Author Responsible for Statistical Analyses: Annouschka Laenen, PhD. Supplementary material associated with this article can be found in the online version at doi:10.1016/j.adro.2026.102072. Contributor Information Julie den Hartog, Email: julie.denhartog@uzleuven.be. Annouschka Laenen, Email: annouschka.laenen@kuleuven.be. Supplementary materials Hematologic toxicity - Suppl Appendix - R3 - clean_V3 References Associated Data Supplementary Materials Hematologic toxicity - Suppl Appendix - R3 - clean_V3