Outcomes of Primary Mediastinal B-cell Lymphoma Patients with Partial Metabolic Response: A Multicenter Retrospective Analysis

Abstract

Purpose

Many patients with primary mediastinal B-cell lymphoma (PMBCL) achieve only a partial metabolic response (PMR) after initial systemic therapy. However, limited data exist on their outcomes. This study aimed to characterize outcomes in patients with PMBCL who achieve PMR and identify factors guiding appropriate treatment for these patients.

Methods and Materials

We reviewed patients PMBCL patients treated at 2 independent cancer centers from January 2009 through September 2021. Using the modified Lugano criteria (2014), end-of-chemotherapy positron emission tomography (PET) scan results were evaluated to assess response. Progression-free survival (PFS) and overall survival (OS) rates from the end-of-chemotherapy PET scan date were estimated using the Kaplan-Meier method.

Results

A total of 151 patients with PMBCL aged between 15 and 65 years were initiated on systemic therapy and underwent a fluorodeoxyglucose PET scan to evaluate response. Of these, 55 (36%) achieved incomplete metabolic response (IMR) (Deauville score [DS] 4 or 5): 13 (8%) progressed on systemic therapy (a DS score of 5), and 42 (27%) achieved a PMR (a DS score of 4). The 4-year PFS and OS rates for all patients (N = 55) with IMR were 73% and 72%, respectively. PMR management included consolidative radiation therapy (RT) in 36 patients (86%), further chemotherapy in 3 patients (7%), and observation in 3 patients (7%). Four-year PFS and OS among all patients with PMR were 83% and 81%, respectively, and 89% and 87% among those receiving RT. Patients with PMR with maximum standard unit value (SUVmax) > 5 had a lower 4-year PFS (74%) compared with those with SUVmax ≤ 5 (95%), although this difference did not achieve statistical significance (P = .07). None of the 3 patients with PMR under observation relapsed.

Conclusions

Patients with PMBCL often have an IMR. PMR (a DS score of 4) managed with subsequent RT is associated with excellent outcomes. SUVmax may identify patients who may require more or less intensive treatment.

Introduction

Primary mediastinal B-cell lymphoma (PMBCL) is a distinct subtype of non-Hodgkin lymphoma (NHL), predominantly affecting young patients.¹ Several studies have indicated that fluorodeoxyglucose (FDG) positron emission tomography (PET) imaging helps in assessing the prognosis for aggressive NHLs, but PET-directed radiation therapy (RT) for PMBCL is less well understood.2, 3, 4, 5 A recent randomized trial (International Extranodal Lymphoma Study Group-37) reported that patients with PMBCL achieving a complete metabolic response (CMR, ie, Deauville scores [DS], 1-3) to initial systemic therapy do not benefit from adjuvant RT.⁵ However, a significant proportion of patients with PMBCL do not achieve a CMR, and less is known about the outcome of those with a partial metabolic response (PMR, ie, DS score 4). The decision to provide RT for these patients requires consideration of the tradeoff between disease control and potential late effects of treatment, with some published data indicating favorable outcomes among patients with a DS score of 4 without further treatment.⁶ This study aimed to evaluate the outcomes of patients with PMBCL with PMR and assess potential prognostic factors.

Methods and Materials

This retrospective study evaluated patients with PMBCL treated at 2 independent cancer centers from January 1, 2009, to September 30, 2021. Patients aged 18 years or older who completed initial immunochemotherapy and whose end-of-chemotherapy PET (EOC-PET) scan showed DS values of 4 or 5 were included. A CMR was defined as a DS score of 1 to 3 on an EOC-PET scan (ie, maximum standard unit value [SUVmax] value of all disease sites less than liver SUVmax value). PMR was defined as a DS score of 4 (SUVmax value greater than liver SUVmax value) without anatomic progression on a computed tomography (CT) scan. Progressive disease was defined as a DS score of 5 with increasing SUVmax values or new disease sites.^7,8

Data were extracted from medical records, including age, gender, presence of B-symptoms, Ann Arbor stage at diagnosis, the Eastern Cooperative Oncology Group performance status score, serum lactate dehydrogenase levels, extranodal involvement, bulky disease (tumor diameter > 10 cm), treatment modalities, and PET-CT scan imaging features (SUVmax values, residual tumor size, and DS scores). Progression-free survival (PFS) and overall survival (OS) from the EOC-PET scan date were calculated using the Kaplan-Meier method. Associations between PFS and various covariates were evaluated using univariate and multivariate Cox proportional hazards models. The prognostic value of SUVmax among patients with PMR was assessed using the median value as the cutoff (SUVmax value = 5). Ethical approval was obtained from the ethics review board at both institutions.

Results

A total of 151 patients with PMBCL underwent an EOC-PET scan, 49 (89%) following rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone 21-day cycle (R-CHOP-21) chemotherapy. Of these, 96 (64%) achieved a CMR, and 55 (36%) had an incomplete metabolic response (IMR) with a DS score of 4 or 5. Among patients with IMR (n = 55), 42 (76%) had a PMR (ie, DS score of 4), and 13 (24%) had refractory or progressive disease (Fig. E1). The median (range) follow-up after the EOC-PET scan of the whole N = 55 cohort was 2.6 (0.2-10.5) years (Table 1).

Table 1.

Baseline demographic and clinical characteristics of the study population

	Metabolic response
Characteristic	Partial response (N = 42)	Progressive (N = 13)	Total (N = 55)
Gender
Male	14 (33.3%)	6 (46.15%)	20 (36.36%)
Female	28 (66.7%)	7 (53.85%)	35 (63.64%)
Age at diagnosis (y)
Mean (±SD)	30.85 (8.9)	33.6 (9.38)	31.50 (8.98)
Median (range)	29.1 (18.4-52.8)	32.1 (21.7-47.8)	29.6 (18.4-52.8)
Stage
I	23 (54.8%)	4 (030.77%)	27 (49.09%)
II	13 (30.9%)	6 (046.15%)	19 (34.55%)
III	1 (2.4%)	0	1 (1.82%)
IV	5 (11.9%)	3 (23.08%)	8 (14.55%)
B symptoms
No	29 (69.05%)	9 (69.23%)	38 (69.09%)
Yes	13 (30.95%)	4 (30.77%)	17 (30.91%)
ECOG performance status score
0	18 (42.86%)	5 (38.46%)	23 (41.82%)
1	18 (42.86%)	5 (38.46%)	23 (41.82%)
2	3 (7.14%)	0	3 (5.45%)
3	2 (4.76%)	3 (23.08%)	5 (9.09%)
4	1 (2.38%)	0	1 (1.82%)
IPI*
0	2 (4.88%)	0	2 (3.70%)
1	25 (60.98%)	8 (61.54%)	33 (61.11%)
2	13 (31.71%)	4 (30.77%)	17 (31.48%)
3	1 (2.44%)	1 (7.69%)	2 (3.70%)*
Large mediastinal mass: 1= yes, 0 = no
No	14 (33.33%)	2 (15.38%)	16 (29.09%)
Yes	28 (66.67%)	11 (84.62%)	39 (70.91%)
Chemotherapy regimen
R-CHOP-21	37 (88.1%)	12 (92.31%)	49 (89.09%)
R-CHOP + DA-R-EPOCH	1 (2.38%)	1 (7.69%)	2 (3.64%)
DA-R-EPOCH	4 (9.52%)	0	4 (7.27%)
EOC biopsy performed	2 (4.8%)	12 (92.3%)	14 (25.5%)
Median SUVmax value median (range)	5.1 (2.6-21.7)	16.3 (7.9-29.8)	6.3 (2.6-29.8)

Abbreviations: DA-R-EPOCH = dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab; ECOG = Eastern Cooperative Oncology Group; EOC = end-of-chemotherapy; IPI = international prognostic index; R-CHOP-21 = rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone 21 days cycle; R-CHOP = rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone; SUVmax = maximum standard unit value.

^⁎IPI missing for 1 patient.

All 13 patients with a DS score of 5 on the EOC-PET scan received salvage chemotherapy, 12 of whom had biopsy confirmation of progressive disease. Of these, 6 (46%) underwent autologous stem cell transplant, while 7 (54%) did not because of progression (N = 5) or comorbidities (N = 2). Compared with patients with PMR (DS score of 4), those with a DS score of 5 had higher progression rates (4-year PFS = 42% vs 83%) and worse 4-year OS (42% vs 81%; P = .009 for PFS, P = .001 for OS; Fig. 1 and Fig. E2).

Figure 1.

Progression-free survival following the end-of-systemic therapy positron emission tomography (PET) scan among patients with an incomplete metabolic response.

Abbreviations: DS = Deauville score.

Among the 42 patients with PMR, 36 (86%) received consolidative involved-site RT, typically 30 Gy to all initially involved lymph nodes with a 5 to 10 Gy boost to the PET-avid volume (median dose to PET-avid site 40 Gy, range, 30-45 Gy). Three patients (7%) were observed with repeat imaging, and 3 (7%) received additional R-CHOP chemotherapy. The 4-year PFS among all patients with PMR was 83% and 89% for those receiving RT (Fig. 2). The 4-year OS was 81% for all patients with PMR and 87% for those receiving RT. One patient treated with RT died of pneumonia without progression.

Figure 2.

Progression-free survival following the end-of-systemic therapy among patients with the partial metabolic response (DS score of 4).

Abbreviations: DS = Deauville score.

Three patients with PMR were managed with observation and repeat PET scans. Their EOC SUVmax values were 2.8, 4.2, and 5.7. After 1 to 3 additional PET scans, 2 patients achieved CMR, while 1 was transitioned to CT imaging without achieving a DS score ≤ 3. No relapses occurred after 20 to 46 months of follow-up (Fig. 2).

Multivariate analysis of data from patients with PMR showed no significant associations between PFS and covariates such as age, gender, stage, bulky disease, IPI score, and serum lactate dehydrogenase levels. Patients with SUVmax values > 5 had a lower 4-year PFS (74%) compared with those with SUVmax values ≤ 5 (95%), although this difference did not achieve statistical significance (P = .07, Fig. 3).

Figure 3.

Progression-free survival among patients with partial metabolic response, stratified using median SUVmax values of end-of-systemic therapy PET scan (SUVmax values ≤ 5 vs > 5).

Abbreviations: PET = positron emission tomography; SUVmax = maximum standard unit value.

*One patient with missing data; 41 patients were included in the analysis.

Discussion

PET scan-guided response assessment is standard for aggressive histology NHLs. The International Extranodal Lymphoma Study Group -37 trial found that patients with PMBCL with CMR do not benefit from adjuvant RT. However, IMR to initial therapy is more common in PMBCL than in other aggressive B-cell lymphomas, and the best way to treat these patients remains unknown, particularly for those who achieve a DS score of 4.

This is one of the largest studies evaluating in detail the treatment and outcomes of patients with PMBCL with PMR. Prior studies have reported 5-year PFS > 85% for patients with DS score 4 who received RT after R-CHOP-21 therapy.^9,10 In a study of 174 patients with PMBCL, Zucca et al¹⁰ found that 86% of patients with a DS score of 4 received RT, and the 3-year PFS was 97.1%, which was comparable to those who achieved CMR.

While RT is commonly used in PMR management, this approach must be balanced against the potential late effects of mediastinal RT, and some studies suggest that not all patients require additional treatment. Melani et al,⁶ for example, reported the outcomes of 25 patients with DS scores of 4 to 5 on EOC-PET scans after DA-EPOCH-R. The crude relapse/progression rate was 20%. The median SUVmax value for patients who did not relapse was 4.7, compared with 15.45 for those who did, suggesting that FDG uptake intensity may help predict prognosis.⁶ Our study, which used limited data on observation-only patients with PMR, found no relapses in 3 patients with a DS score of 4 and low SUVmax values monitored with recurrent PET scans. The majority of patients with PMR in our study, however, responded well to RT, and given the poor prognosis of relapsed PMBCL, the criteria for selecting observation-only patients should be studied in clinical trials.

This study has limitations, including potential selection biases and unmeasured confounders. The small number of patients in some subgroups limited our ability to evaluate prognostic biomarkers or adjust for selection bias among treatment subgroups. Most patients received R-CHOP-21 therapy, and there are persuasive data that more intensive regimens (eg, R-CHOP 14-day cycle therapy) can produce higher rates of CMR and better PFS for patients with PMBCL.^11,12

Conclusions

Patients with PMBCL who achieved a PMR through initial immunochemotherapy had an excellent prognosis with consolidative RT. However, the prognosis remains poor among those whose disease is refractory or progressive. SUVmax values (or other measures of FDG uptake) may help identify patients who could be managed with observation. However, considering the poor outcome of relapsed PMBCL, clinical trials should be conducted to determine whether or not to provide RT to patients with PMR.

Disclosures

Ur Metser reports financial support (institutional fund) from the INOVAIT grant and consulting fees from Radialis. Ur Metser also reports serving as a leader for the chair of the PET steering committee, OH-CCO. Anca Prica reports receiving honoraria from AbbVie, AstraZeneca, and Kite-Gilead. Michael Crump reports financial support (institutional fund) from Epyzyme/Ipsen clinical trials funding and Roche Canada clinical trials funding and consulting fees from Canada's Drug Agency (formerly CADTH) and Kyte Gilead. Danielle Rodin reports financial support (institutional fund) from the Leukemia and Lymphoma Society and the Canadian Cancer Society. Danielle Rodin reports serving as a clinical guidelines committee member for Ontario Health. Other authors have nothing to disclose.