A single-center,retrospective analysis of relapse and progression patterns of pr

QIN Yue ,LIU Rongping ,ZHANG Xiaonan ,ZHANG Wan ,REN Chen ,WU Dehua

1Department of Radiation Oncology,Nanfang Hospital,Southern Medical University,Guangzhou, 510515,China;2Department of Radiation Oncology,Sun Yat-sen University Cancer Center,Guangzhou 510060,China

Abstract: Objective To analyze recurrence and progression patterns of primary central nervous system lymphoma (PCNSL) in patients without whole brain radiotherapy(WBRT)and assess the value of WBRT in PCNSL treatment.Methods This retrospective single-center study included 27 patients with PCNSL,who experienced recurrence/progression after achieving complete remission(CR),partial remission,or stable disease following initial treatments with chemotherapy but without WBRT.The patients were followed up regularly after the treatment for treatment efficacy assessment.By comparing the anatomical location of the lesions on magnetic resonance images (MRI) at the initial diagnosis and at recurrence/progression,we analyzed the patterns of relapse/progression in patients with different treatment responses and different initial status of the lesions.Results MRI data showed that in 16(59.26%)of the 27 patients,recurrence/progression occurred in out-field area(outside the simulated clinical target volume[CTV])but within the simulated WBRT target area in 16(59.26%)patients,and within the CTV(in-field)in 11(40.74%)patients.None of the patients had extracranial recurrence of the tumor.Of the 11 patients who achieved CR after the initial treatments,9(81.82%)had PCNSL recurrences in the out-field area but within WBRT target area;of the 13 patients with a single lesion at the initial treatment,11 (84.62%) experienced PCNSL recurrence in the out-field area but within WBRT target area.Conclusions Systemic therapy combined with WBRT still remains the standard treatment for PCNSL patients,especially those who achieve CR after treatment or have a single initial lesion.Future prospective studies with larger sample sizes are needed to further explore the role of low-dose WBRT in PCNSLtreatment.

Keywords: primary central nervous system lymphoma,radiation therapy,whole brain radiotherapy,recurrence/progression pattern

INTRODUCTION

Primary central nervous system lymphoma (PCNSL) is a rare subtype of non-Hodgkin lymphoma that typically affects the brain,soft meninges,spinal cord,and the eyes［1］,and accounts for approximately 4% of primary brain or CNS tumors and 5%-8% of extranodal lymphomas［2,3］.Although PCNSL is sensitive to both chemotherapy and radiotherapy,the overall response rates (ORR) and long-term survival of the patients are poor compared to similar subtypes of extranodal NHL,with 5-year survival rates of 20%-30%［4］.

Whole brain radiotherapy (WBRT) combined with glucocorticoids used to remain the standard treatment for PCNSL,but given the risk of treatment-related neurotoxicity,particularly in the elderly,there has been a shift towards alternative consolidation strategies.Highdose methotrexate-based (HD-MTX) chemotherapy is now considered the cornerstone of treatment for PCNSL［5,6］.Nonetheless,relapse occurs in more than 50%of patients within two years after the treatment［7］.

The Radiation Therapy Oncology Group (RTOG)demonstrated that a combined therapeutic approach with WBRT extended the mean progression-free survival(PFS) of patients with PCNSL to 104.3 months［8］.Rubenstein JL et al［9］were the first to propose a consolidation strategy that did not involve WBRT,using HD-MTX,temozolomide,and rituximab (R) induction chemotherapy followed by consolidation therapy with etoposide plus cytarabine in patients who achieved complete remission (CR),and they concluded that this regimen was comparable to a WBRT-involving regimen in terms of the PFS.Autologous stem cell transplantation(ASCT)-supported high-dose chemotherapy is the most studied alternative to WBRT as a consolidation therapy［10］,but its impact on cognitive functionis stillunclear.

The role of WBRT in PCNSL treatment remains debatable.In this study,we conducted a retrospective analysis of patients with recurrent or progressive PCNSL who had not previously undergone WBRT.Simulated target delineation was performed to compare the recurrent lesions with the primary tumor and radiotherapy target areas.We analyzed the recurrence or progression patterns of PCNSL to assess whether WBRT could be replaced and the potential value of WBRT in reducing PCNSL recurrence following successful initial treatments.

METHODS

Patients

We reviewed the data of patients with newly diagnosed PCNSL at our hospital between January,2009 and December,2020 using the Medical Big Data Platform of our hospital.Inclusion criteria:(1) patients with a histologically confirmed diagnosis of PCNSL at the time of initial treatment without manifestations of lymphoma at sites other than the CNS;(2)patients who did not receive WBRT in the initial treatment and experienced relapse or progression after achieving CR,partial remission(PR),or stable disease (SD) during or after the treatment;and (3)patients with complete magnetic resonance imaging(MRI) data both at the initial diagnosis and for response evaluation.Exclusion criteria:(1)patients with lesions in the eye;(2) patients whose MRI data were not evaluable for efficacy assessment.

A total of 27 patients were included,and the data were collected including the patient's age at diagnosis,gender,Karnofsky (KPS) score,site of primary tumor lesion,number of primary tumor lesions,treatments received prior to recurrence or progression,efficacy evaluation,site of tumor recurrence,and the time from diagnosis to imaging assessment of recurrence or progression.This study was approved by the Ethics Committee of Nanfang Hospital,Southern Medical University with waiver of informed consent.

Imaging and treatment response assessment

All the patients included in this study were first analyzed for anatomical location of PCNSL by cranial contrast-enhanced MRI or cranial contrast-enhanced CT,and the lesions elsewhere were excluded by PET/CT or thoracoabdominal CT.CT images with high signal intensity at the initial diagnosis as well as enhanced T1-weighted MR images were collected.

Treatment response was assessed using the treatment efficacy assessment criteria established by the International PCNSL Collaborative Group (IPCG)［11］.Efficacy assessment was performed approximately every 2 months during the treatment or when the treatment modality was changed.The patients were followed up regularly after the end of treatment,and efficacy assessment was performed every 3 months in the first 2 years,every 6 months at 3-5 years,and annually thereafter［12］.

At the time of recurrence,cranial contrast-enhanced MRI was used for tumor assessment.According to the location of the enhanced lesions on T1-weighted MRI images,CNS lesions can be further divided into the following sites:(A) the same lobe;(B) the same lobe but distal (in the same cranial lobe in the contralateral cerebral hemisphere);(C) different lobes (in different cranial lobes in the ipsilateral and contralateral cerebral hemispheres);(D) the corpus callosum;(E) the lateral ventricles;(F) the basal ganglia region;(G) the infratentorial posterior fossa (the brainstem and the cerebellum;and (H) the centrum semiovale.All MRI examinations were reassessed by experienced neuroradiologists.

Definition of recurrence/progression pattern

As PCNSL is a diffusely infiltrative disease without a consensual definition of its margins,we defined its recurrence/progression pattern in this study based on the simulated margins used to specify the radiation treatment plans for glioblastomas［13］.As shown in Fig.1,the simulated gross tumor volume (GTV) was the lesion delineated on an enhanced T1-weighted MRI image at the time of the initial treatment,and the simulated clinical target volume (CTV) was outlined by adding a 2 cm margin to the GTV.The site of recurrence or progression,defined by the position of the recurrent lesion relative to the simulated local target and WBRT target,was divided into:(1) in-field:recurrence or progression lesions in the simulated CTV;(2) out-field but within the simulated WBRT target area:recurrence or progression lesions outside the simulated CTV but within the brain;(3) out of the brain:recurrence or progression lesions outside the cranium.

Fig.1 Schematic representation of recurrence or progression patterns of PCNSL.1:In-field recurrence;2:Out-field recurrence but within the simulated WBRT target area;3:Out-field recurrence(outside of the brain).The simulated GTV is shown in red and the simulated CTV in white.

RESULTS

Clinical characteristics of the patients

We initially identified 235 patients with a possible diagnosis of PCNSL,and 190 patients were excluded for absence of diagnosticdata,unknown diagnosis,previous WBRT,or absence of evaluable MRI data (Fig.2).Forty-five patients with histologically confirmed PCNSL were finally screened,and 27 of them experienced relapse or progression during the follow-up,with a median PFS of 20.0 months (8.8-31.2 months;Fig.3).Sixteen of the 27 patients were female (16/27,59.26%)and 11 were male(11/27,40.74%),with a median age of 53 years (23-72 years) and a median PFS of 10 months(95%CI:6.9-13.1 months).Eighteen of the patients had a KPS ≥70,5 had a KPS＜70,and KPS was not available in 4 patients.The lesions were confirmed pathologically by stereotactic biopsy in 18 cases,by surgery in 8 cases,and by endoscopic biopsy in 1 case.Thirteen(48.15%)of the cases had a single lesion at the initial diagnosis,and 14(51.85%) had multiple lesions;16 cases (55.56%) had unilateral lesions,and 11(40.74%)had bilateral lesions.Of the 27 patients with tumor recurrence,the initial treatments resulted in CR in 11 patients,PR in 15 patients,and SD in 1 patient.The recurrent/progression sites involved the same lobe in 6 cases,different lobes in 5 cases,the basal ganglia in 5 cases,the same lobe but at a distant site in 2 cases,the infratentorial posterior fossa in 3 cases,the lateral ventricle in 3 cases,and the centrum semiovale in 1 case,the cerebral hemisphere in the 1 case,and the corpus callosum in 1 case(Tab.1).

Fig.2 Screening procedures for PCNSL patients with recurrence or progression.

Fig.3 Cumulative survival rate of 45 patients with PCNSL.

Tab.1 Demographic and clinical data of the 27 patients with PCNSL relapse

Recurrence or progression sites of PCNSL

In the 27 patients,the recurrence/progression sites were out-field but within simulated WBRT target area in 16(59.26%) patients,and among them the lesions involved different lobes in 5 (5/16,31.25%) cases,the lateral ventricle in 3(3/16,18.75%)cases,and the infratentorial posterior fossa in 3 (3/16,18.75%) cases;in the remaining 5 cases (5/16,31.25%),the recurrence/progression sites involved the same lobe but at a distant site,the basal ganglia region,the centrum semiovale and the cerebral hemisphere (Fig.4A).Eleven (11/27,40.74%) patients had in-field recurrence/progression,involving the same lobe in 6 (6/11,54.54%),the basal ganglia region in 4 cases(4/11,36.36%),and the corpus callosum in 1 case (1/11,9.10%) (Fig.4B).None of the patients had extracranial recurrence.

Recurrence/progression patterns in patients with different primary lesions

Thirteen patients had single lesions at the initial treatment,and among them 11 (84.62%) experienced recurrence at out-field sites but within simulated WBRT target area.Of these 11 patients,relapse occurred in different lobes in 4(36.36%)patients,in the same lobe in 2 (18.18%) patients,in the basal ganglia region in 1(9.10%) patient,in the lateral ventricles in 1 (9.10%)patient and in the infratentorial posterior fossa in 1(9.10%)patient.Among the 13 patients,4 received gross total resection (GTR) and HD-MTX with or without R,2 received GTR alone and 2 received HD-MTX chemotherapy alone combined with stereotactic radiosurgery (SRS)/gamma knife radiosurgery (GKS) and HD-MTX with or without R,and 1 received GTR and HD-MTX with R and immunotherapy.The remaining 2 patients had in-field progression,occurring in the basal ganglia in 1 case and the infratentorial posterior fossa in the other(Fig.4E,Fig.5C);one patient received SRS/GKS combined with HD-MTX,and the other received GTR and HD-MTX combined with R and cytarabine.

Fourteen patients had multiple lesions at the initial treatment,and out-field progression but within simulated WBRT target area occurred in 5 (35.71%) cases involving different lobes,the infratentorial posterior fossa,the centrum semiovale,the basal ganglia region,or the same lobe.Four out of the 5 patients received HD-MTX chemotherapy alone,and 1 (1/5,20%) patient received SRS/GKS combined with HD-MTX.In-field occurrence was found in 9 patients (9/14,64.29%),involving the same lobe in 5(5/9,55.56%)cases;1(1/9,11.11%)patient had primary focal progression,and 3(3/9,33.33%) had recurrence in the basal ganglia (Fig.4F and Fig.5D).Of these patients with in-field recurrence,5 received HD-MTX chemotherapy alone and 4 received SRS/GKS combined with HD-MTX.

Relapse/progression patterns in patients with different treatment responses

Eleven (40.74%) of the 27 patients achieved CR after treatment,and 9 (9/11,81.82%) of them had out-field recurrence but within the brain,including 4 (4/9,44.45%)with recurrence involving different lobes,2(2/9,22.22%)with recurrence in the lateral ventricles,2 in the infratentorial posterior fossa,and 1 (1/9,11.11%) in the basal ganglia region (Fig.4C,Fig.5A).The initial treatment modalities in the 9 patients included GTR and HD-MTX in 4 patients,GTR alone in 2 patients,HD-MTX chemotherapy alone in 2 patients,and GTR,HD-MTX combined R and immunotherapy in 1 patient.Two of the remaining 11 patients (2/11,18.18%) had local in-field recurrence and received SRS/GKS combined with HD-MTX,GTR and HD-MTX combined with R and cytarabine,respectively.

Fifteen(15/27,55.56%)patients achieved PR after the initial treatment,and 7 (7/15,46.67%) of them experienced out-field progression in scattered locations within the brain,involving the same lobes but at a distant site,different lobes,the infratentorial posterior fossa,the basal ganglia region,the centrum semiovale,and the lateral ventricle.Three of the patients received SRS/GKS and HD-MTX with or without R,and 4 received HD-MTX chemotherapy alone.In-field progression occurred in 8 cases (8/15,53.33%),involving the same lobe in 3 (3/8,37.5%)cases,the basal ganglia region in 4(4/8,50.00%)cases,and the corpus callosum in 1 case (1/8,12.5%)(Fig.4D,Fig.5B).In terms of treatment regimen,4 patients received HD-MTX chemotherapy alone and the other 4 received SRS/GKS,HD-MTX with or without R.

Fig.4 Sites of recurrence or progression of PCNSL in patients with out-field(but within the brain)recurrence(A),in-field recurrence (B),complete remission (C),partial remission (D),single primary lesions (E),and multiple primary lesions(F).

Fig.5 Pie charts showing percentages of in-field and out-field(within the brain)recurrence in patient subgroups with complete remission (A),partial remission(B),single primary lesion(C),and multiple primary lesions(D).

Imaging evaluation of patients with relapse after treatment up to CR

Due to the small sample size,we did not made statistical comparisons among the subgroups and selected 2 representative cases for description.

Case No.16 was a 54-year-old male patient with a KPS score ＞70.At the initial diagnosis,contrastenhanced T1-weighted MR images on cross-section showed a lesion in the right frontal lobe (Fig.6A).The patient underwent gross total resection (GTR) of the tumor and its surrounding edematous tissue,which was confirmed microscopically (Fig.6B).After surgery,the patient received 4 cycles of R combined with HD-MTX therapy and was monitored regularly.Follow-up MRI at 21 months revealed multiple round shadows in the right frontal area and the right temporoparietal-occipital junction area(Fig.6C),and the largest lesion was 4.5 cm×3.4 cm.

Case No.225 was a 63-year-old female patient with a KPS score ＜70.Contrast-enhanced T1-weighted MR image at the initial diagnosis showed a single lesion(4.1 cm×4.2 cm×4.4 cm) in the right frontal falciform area(Fig.6D).Postoperative MRI showed no evidence of residual tumor and the patient was evaluated to have a complete response (Fig.6E).The patient received 6 cycles of HD-MTX chemotherapy after surgery.Two months after completion of chemotherapy,follow-up MRI showed tumor recurrence in the left medipeduncle(8 mm),the left cerebellar hemisphere (7 mm),and the left side of the dorsal medulla(1.2 cm×0.9 cm)(Fig.6F).

Fig.6 MR images of a 54-year-old male patient (Case No.16;A-C) and a 63-year-old female patient (Case No.225;D-F) with PCNSL recurrence.A:At the initial diagnosis,a T1-weighted image showed an isolated lesion in the right medial frontal lobe(arrow).B:MRI showed no visible lesions after complete remission(CR)following gross total resection(GTR).C:At 23 months after achieving CR,MRI showed local recurrence in the operation region and distant recurrence in the right parietal/temporal/occipital junction(arrows).D:At the initial diagnosis,a T1-weighted image showed an isolated lesion in the right medial frontal lobe(arrow).E:After achieving CR following GTR,no visible lesions were found.F:At 6 months after achieving CR,MRI revealed distant recurrence in the medulla oblongata dorsal(arrow).

DISCUSSION

In this study,we evaluated the recurrent/progression sites of PCNSL in patients who did not receive WBRT and compared them with the simulated radiotherapy target area to evaluate the potential effect of WBRT for PCNSL control.Our data showed that among the PCNSL patients who experienced relapse or progression after achieving CR following first-line treatment without WBRT,81.82% had recurrence at the out-field sites but within the brain.In PCNSL patients with a single initial lesion,84.62%had recurrence or progression in out-field area but within the brain.These patients all underwent systemic therapy and local aggressive treatment,but tumor recurrence still occurred after the local lesions were well controlled,and most of the patients had recurrence in the out-field area but within the brain.This seems to suggest that consolidation therapy with WBRT is still needed in patients who achieve CR after treatment or have a single initial lesion.Our finding are generally consistent with the results of previous studies［14-16］.In most of the cases of PCNSL relapse,recurrences occur within the CNS and within the irradiation range of WBRT target area,and only a few cases had recurrence in the more distal sites outside the CNS,such as the lungs,testes,and adrenal glands［17］.Consolidation therapy with WBRT after initial treatment is likely to reduce tumor recurrence in most of the cases.

Our results also showed that the patients with initial single lesions who had CR after treatment tended to experience recurrence at a different anatomical site other than the primary site.This finding is consistent with the results of Schulte-Altedorneburg et al［14］and Tabouret et al［18］.Furthermore,we found that the site of recurrence often involved a different and distant cranial lobe,suggesting the presence of unknown mechanisms of recurrence other than local infiltration.These results further support the importance of WBRT in patients receiving enhanced systemic therapy and local treatment for the primary lesion.

Since the 1980s,WBRT has been a standard radiation therapy for patients with PCNSL,but concerns remain over the high risk of delayed neurotoxic reactions associated with WBRT［19］.Attempts have been made to replace radiotherapy with other alternatives including ASCT,but according to Ferreri et al［20］,who reported the long-term efficacy and neural tolerability of the IELG32 trial,ASCT resulted in little difference in 7-year OS of the patients;among the patients who achieved partial response after induction,the 7-year PFS was even higher in WBRT group than in ASCT group (54%vs46%,P=0.43),and the MMSE did not differ significantly between the two groups.Therefore,WBRT can not be completely replaced by ASCT.

Reduced-dose WBRT (rdWBRT) is suggested to provide a more effective resolution［21］.In 2013,Morris et al［22］proposed chemotherapy combined with consolidation rdWBRT for treatment of primary CNS lymphoma,and this reduced-dose radiotherapy regimen has been widely accepted and recommended.But in a prospective phase II trial of this regimen by Adhikari et al［23］,tumor relapse still occurred in 3 patients following rdWBRT,while no relapses or disease-related deaths occurred in patients receiving standard-dose WBRT.Wu et al［24］further proposed to prescribe different radiation doses for patients aged 60 years or below who achieved CR and for those who did not achieve CR regardless of age,and the results showed an ORR of 90.9% in 11 patients who received salvage WBRT.These findings show that either rdWBRT or ASCT can improve the outcomes of young PCNSL patients,while salvage WBRT can provide ORR benefit with good tolerability in older patients with poor responses to chemotherapy.

Conclusion

Our findings show that in PCNSL patients who experience relapse after first-line treatment without WBRT,the recurrent lesions are often located in the out-field areas but confined within the simulated WBRT target area.This highlights the importance of WBRT in both the induction and consolidation stages of treatment.Furthermore,personalized rdWBRT tailored to different patients and different treatment responses may help to minimize the neurotoxicity.Given the retrospective nature and the small sample size of this study,these findings await validation by future multicenter,prospective,randomized controlled trials.