Imaging Surveillance of Limited-Stage Classic Hodgkin Lymphoma Patients after PET-CT Documented First Remission

Abstract

Purpose/Objective:

Early stage Hodgkin lymphoma (ESHL) is highly curable, however 10–15% of patients experience relapse. We examined the utilization of follow-up imaging for ESHL patients who achieved a metabolic complete response after upfront therapy.

Materials/Methods:

The records of adult patients treated at a single institution between 2003 and 2014 were reviewed. Positron emission tomography-computed tomography (PET-CT) and CT scan frequency was quantified during the 2 years following treatment and subsequent visits beyond 2 years.

Results:

The study cohort contained 179 patients. The median age was 31 years; bulky disease was present in 30%. Doxorubicin, bleomycin, vinblastine and dacarbazine (ABVD) or AVD was given in 97%; 75% received radiation therapy. At a median follow up of 6.9 years, the 5-year progression-free (PFS) and overall survival (OS) rates were 93.7% and 98.1% respectively. Relapse occurred in 5% of patients (n=9) at a median of 9.1 months (range 4.6 – 27.2) from therapy. Two patients presented with symptoms prompting imaging in follow-up. Within 2 years after therapy, 376 PET-CT scans and 3325 CT scans were performed, yielding an average of 2.1 PET-CTs and 18.6 CTs per patient. Of the initial 179 patients, 113 had follow up conducted beyond 2 years post-therapy; an average of 2.7 PET-CTs and 33.2 CTs were performed. In the 2 year post therapy period, 463 scans were performed per relapse detected.

Conclusion:

In this cohort of ESHL patients who responded completely to frontline therapy, the relapse rate was low. Routine imaging surveillance lacks clinical benefit in this patient population.

Introduction

Limited-stage (stage I or II) classical Hodgkin lymphoma is a highly curable disease with 5-year overall survival rates in excess of 90%.^1,2 However, approximately 10–15% of patients experience a relapse after complete response to therapy, but over half will go on to achieve excellent disease-free survival with salvage therapy.^3,4,5

Identifying relapsed disease is important so that salvage therapy can be initiated. Patients are at the greatest risk of relapse within two years of therapy completion.³ For this reason, the most intensive follow up for disease recurrence is in the 2 year window after completion of treatment. Controversy exists regarding the optimal imaging surveillance strategy^6,7. The most recent National Comprehensive Cancer Network (NCCN) imaging follow up schedule suggests computed tomography (CT) imaging of the neck, chest, abdomen and pelvis at 6, 12 and 24 months following the end of therapy is acceptable practice.⁸ Other guidelines strongly discourage surveillance CT scans after documentation of remission⁹. Given the low positive predictive value of positron emission tomography-computed tomography (PET-CT) imaging in the post-therapy period, there is even less enthusiasm for routine PET-CT imaging for surveillance^10–12. This practice is discouraged by the NCCN as well as European guidelines.^13,14

Despite the growing body of literature highlighting the limited utility of routine imaging surveillance among the early stage HL population, oncologists and patients may be hesitant to forgo these scans in clinical practice. We sought to evaluate the impact of surveillance CT and PET-CT imaging among stage I/II patients that achieved a complete metabolic response to treatment with doxorubicin, bleomycin, vinblastine and dacarbazine (ABVD) with or without consolidative radiation therapy (RT).

Methods

Patient Selection/ Inclusion Criteria

After approval by our institutional review board, we retrospectively reviewed the records of all patients with a diagnosis of limited stage HL that were treated at our institution between 2003 and 2014. Patients ≥ 18 years with biopsy-proven Ann Arbor stage I or II disease that achieved a metabolic complete response to frontline therapy, chemotherapy alone or combined modality therapy –(defined as Lugano five point scale (5PS) score of 1–3 on end of therapy PET-CT) were included⁹. The initial screen yielded 267 patients. Patients were excluded from the analysis if they had had follow up for less than 2 years at our institution (n=65), primary refractory disease (n=18), died during ABVD chemotherapy (due to septic shock [n=1] or bleomycin related respiratory failure [n=2])) or died from causes unrelated to HL within the two year follow up period (n=2) (Figure 1). The final cohort was comprised of 179 patients.

Figure 1.

Details on eligible patients.

Patient, Disease and Treatment Characteristics

Patient, disease and treatment characteristics were obtained from the electronic medical record. Disease was staged according to the Ann Arbor system⁹. Patients were categorized into early-stage favorable, early-stage unfavorable, or advanced groups based on the German Hodgkin Study Group (GHSG) risk stratification^15,16. Consistent with GHSG groupings, stage IIB patients with bulky disease or extranodal disease were classified as advanced. Bulky disease was defined as a mediastinal nodal mass or conglomerate >10 cm in the axial, sagittal, or coronal dimensions.

Treatment generally consisted of ABVD with or without consolidative RT. PET-CT imaging was performed either during or at the end of systemic therapy to confirm complete response. The 5PS was used to determine response to therapy with 5PS of 1–3 considered negative⁹. Patients with primary refractory disease that received salvage therapywere not included in the study. The end of treatment date was defined as the date of the final restaging radiographic study at the end of chemotherapy or in cases where RT was given, the end of RT. The type of imaging utilized for surveillance as well as the frequency of the follow up imaging was at the discretion of the treating physician. Clinical surveillance was not employed.

Surveillance Imaging Quantification and Radiation Dose Calculations

Scan numbers for PET-CT scans and CT scans were retrieved from the electronic medical record and quantified for two years post-treatment for all 179 patients. CT scans of the neck, chest, abdomen, and pelvis were counted separately with each type of anatomical scan counting as “1” scan. For 113 patients that had long-term follow up at our institution, a “lifetime” post therapy scan count was performed. Radiation exposure secondary to imaging studies was calculated based on standard institutional CT protocols (CT of the neck, 2–3.5 mSv; chest, 3.3–3.7 mSv and abdomen/pelvis combined, 13–15 mSv). The higher end of the range was used for calculations. For PET-CT imaging the total exposure based on institutional protocols is roughly 12 mSv (7 mSv from 10mCi of ¹⁸F-fluorodeoxyglucose [FDG] and 5 mSv from the CT scan). For calculations of lifetime excess cancer risk secondary to imaging studies, 1 Sv exposure was assumed to result in excess lifetime risk of 5% for the general population^17,18.

Statistical Analysis

Progression free survival (PFS) was defined as disease relapse or death from any cause and was calculated from the date of diagnosis. Probability of PFS and overall survival (OS) were estimated with the Kaplan-Meier method¹⁹. When comparing the impact of PET-CT surveillance on OS, for patients with relapse the PET-CT count only included scans ordered for routine surveillance prior to salvage therapy. The log rank test was used for comparison of survival curves.

Results

Baseline patient and treatment characteristics are depicted in Tables 1 and 2. Stage II HL was present in 83% of patients; 30% had bulky disease. ABVD was given with (89%) or without (8%) bleomycin in the majority of patients. One elderly patient was treated with RT alone due to co-morbidities. Consolidative RT was administered in 75% of patients.

Table 1:

Patient Characteristics

Characteristics	All Patients, n=179 (%)	Patients without relapse, n=170 (%)	Patients with relapse, n=9 (%)
Age at diagnosis
median	31	31	38
range	18–87	18–87	18–60
% of 45 and older	43 (24)	41 (24)	2 (22)
Female	105 (59)	101 (59)	4 (44)
Race
Asian	9 (5)	8 (5)	1 (11)
Black	7 (4)	7 (4)	0
Hispanic	24(13)	23(14)	1 (11)
Unknown	8 (4)	7 (4)	1 (11)
White	131 (73)	125 (74)	6 (67)
Stage
I	30 (17)	29 (17)	1 (11)
II	149 (83)	141 (83)	8 (89)
B symptoms	39 (22)	35 (21)	4 (44)
Histological subtype
Nodular sclerosing	143 (80)	136 (80)	7 (78)
Other	36 (20)	34 (20)	2 (22)
Extranodal disease	7 (4)	7 (4)	0 (0)
Total number of Ann Arbor nodal regions/sites
1	29 (16)	28 (16)	1 (11)
2	42(23)	39(23)	3 (33)
3	45 (25)	43 (25)	2 (22)
4+	41 23)	39 (23)	2 (22)
GHSG classification
Early-favorable	54 (30)	52 (31)	2 (22)
Early-unfavorable	120 (67)	114 (67)	6 (67)
Advanced	5 (3)	4 (2)	1 (11)
Bulky Disease	54 (30)	50 (29)	4 (44)

Abbreviations: GHSG, German Hodgkin Study Group

Table 2:

Treatment Characteristics

Treatment Characteristics	All Patients, n=170 (%)	Patients without Relapse, n=170 (%)	Patients with Relapse, n=9 (%)
Chemotherapy regimen
ABVD	159 (89)	151 (89)	8 (89)
AVD	14 (8)	13 (8)	1 (11)
other	5 (3)	5 (3)	0
none	1 (1)	1 (1)	0
Chemotherapy cycle #
median	5	5	6
range	0–6	0–6	3–6
6 cycles	77 (43)	69 (41)	8 (89)
Consolidative RT	135 (75)	132 (78)	3 (33)
RT dose
median	30.6	30.6	30.6
range	20–40	20–40	30.6–36

Abbreviations: ABVD, doxorubicin, bleomycin, vinblastine and dacarbazine; AVD, doxorubicin, vinblastine and dacarbazine; RT, radiation therapy

Among the 44 patients treated with systemic therapy alone, response assessment was based on a PET-CT performed upon completion of systemic therapy among 32 patients and based on a PET-CT scan that was conducted during systemic therapy in 8 patients. In this group of patients treated with systemic therapy alone, the 5PS was 1–3 in 41 patients and 5PS of 4 in 3 patients. Among the 3 patients with 5PS of 4 it was interpreted by the radiologist as treated disease or residual thymic tissue and no further treatment was administered.

Combined modality therapy was administered to 135 patients. In the combined modality group 110 patients had PET-CT imaging performed upon completion of all systemic therapy but prior to RT and 25 patients had PET-CT imaging performed during chemotherapy and not repeated again just prior to RT. Among this entire group of 135 patients, the 5PS was 1–3 in 126 patients and 4–5 in 9 patients. Among the patients with 5PS of 4–5 the PET-CT findings were interpreted as inflammation versus residual disease and biopsy was not performed.

At a median follow up of 6.9 years, the 5-year progression-free (PFS) and overall survival (OS) rates were 93.7% and 98.1% respectively (Figure 2a, 2b). Nine relapses occurred at a median time to relapse of 9.1 months (range 4.6 – 27.2) from end of treatment; all but one relapse (27.2 months) occurred 2 years or earlier from completion of all therapy (Table 3). Two patients presented with symptoms (palpable neck adenopathy) that prompted imaging that led to relapse detection. The remaining 7 relapses were detected on routine surveillance imaging. Five relapses were suspected based on routine surveillance CT imaging that prompted PET-CT scans (n=3) or were seen on both follow up contrast-enhanced CT and non-contrast-enhanced PET-CT scans that were performed on the same day (n=2). The remaining 2 relapses were detected on surveillance PET-CT imaging. The median size of the site of nodal disease detected on imaging was 2.9 cm (range 1.1 – 4.7 cm). The 5-year OS for patients that had PET-CT surveillance imaging within the 2-year post-treatment period was 98.5% (n=161) compared to 94.4% for patients that did not ( n=18, p=0.749, Figure 2c).

Figure 2.

Progression free survival (a) overall survival (b) and overall survival according to surveillance strategy (c) among 179 limited stage Hodgkin lymphoma patients. Dotted lines reflect 95% confidence interval limits.

Table 3.

Details of patient relapses

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6	Patient 7	Patient 8	Patient 9
Age at diagnosis	54	21	38	42	24	21	43	60	18
Gender	Male	Male	Male	Male	Female	Female	Female	Female	Male
Stage	II	I	II	II	II	II	II	II	II
Bulky	No	Yes	No	No	Yes	No	Yes	Yes	No
Chemotherapy	ABVD	ABVD	ABVD	ABVD	ABVD	ABVD	ABVD	ABVD	AVD
Chemotherapy cycle #	6	6	3	6	6	6	6	6	6
Radiation Therapy	No	Yes	Yes	No	No	No	No	Yes	No
End of the therapy PET-CT assessment*	3	1	2	3	3	3	2	3	3
Relapse Detection	Clinical	CT	CT & PET-CT	CT	CT	PET-CT	Clinical	CT & PET-CT	PET-CT
Symptomatic at time of relapse	Yes	No	No	No	No	No	Yes	No	No
Symptom	Palpable neck adenopathy						Palpable neck adenopathy
Physical exam finding at time of relapse	Yes	No	No	No	No	No	Yes	No	No
Location of relapse	above and below diaphragm	mediastinum	mediastinum	mediastinum	mediastinum	mediastinum	neck	above and below diaphragm	above and below diaphragm
Maximum size of relapsed disease (cm)	2	3.2	1.8	3.4	4.7	3	2.9	1.1	1.7
Time from end of treatment to relapse (months)	27.2	7.7	21.5	5.4	11.2	9.1	4.6	5.7	22.9

Abbreviations: PET-CT, positron emission tomography-computed tomography; CT, computed tomography

^*For patients that received combined modality therapy, the end of therapy PET-CT assessment was prior to radiation therapy.

During the two years after treatment completion for all 179 patients, an average of 2.1 PET-CTs and 18.6 CTs were performed per patient (Table 4). CT scans of the neck, chest, abdomen, and pelvis were counted separately with each type of anatomical scan counting as one scan even If the scans were performed during a single session. For the 170 patients that did not experience relapse of HL, an average of 2.0 PET-CTs and 18.7 CTs per patient were conducted. Of the initial 179 patients, 113 patients had follow up conducted at our institution beyond 2 years post-therapy. Among these patients, an average of 2.7 PET-CTs and 33.2 CTs were performed per patient.

Table 4:

Quantification of Imaging Studies

	Scan Count within 2 years post-treatment (n=179)	Life-Time Scan Count Post-Treatment (n=113)
PET-CT	376	302
CT total	3325	3756
neck	799	874
chest	865	974
abdomen	835	951
pelvis	823	941
other	3	16
Total Scan Count	3701	4058

Abbreviations: PET-CT, positron emission tomography-computed tomography; CT, computed tomography

For the 2 years following therapy completion, the average cumulative dose of radiation exposure due to surveillance imaging was 128.2 mSV, of which diagnostic CT scans contributed most of the dose (103 mSv, 80.4%). This corresponds to an estimated lifetime excess cancer risk of 0.6%. In the 2 year post therapy period only 0.2% of surveillance scans resulted in a diagnosis of relapse. There were 8 relapses identified in the 2 year post therapy interval, which corresponds to a number needed-to-scan/treat (NNT) of 463 scans to detect one relapse. For the 113 patients followed past 2 years, the average cumulative dose of exposure was 216.6 mSv. Similarly, diagnostic CT scans contributed the majority of the dose (184.5 mSv, 85.2%). The estimated lifetime excess cancer risk among this group of patients with more extended follow up was 1.1%.

In the group of 126 patients that were treated with combined modality therapy and had a negative PET-CT scan (5PS of 1–3) prior to initiation of RT (PET-CT performed either during systemic therapy and prior to RT [n=25] or upon completion of all chemotherapy and just prior to RT [n=101]), there were only 3 relapses. Additional surveillance PET-CT scans after RT were performed in 115 of 126 patients (91%). The 3 relapses that occurred in this group were detected by either CT alone (n=1) or both PET-CT and CT scans performed during the same follow up visit (n=2). Among the 9 patients with a PET-CT scan with a 5 PS of 4 (n=7) or 5PS of 5 (n=2) after completion of systemic therapy but prior to RT, additional PET-CT imaging was performed after RT in all patients. Subsequent PET-CT scans in this group of patients were interpreted as negative and no relapses occurred.

The number of scans performed within the 2 year period following treatment completion per patient decreased over time. For patients diagnosed in 2003 the average number of CT scans performed was 23.5, while in 2014 the average number of CT scans performed per patient was 12.1 (Figure 3). An average number of 12 scans per patient is consistent with 3 sets of imaging studies performed over the 2 year period with CT imaging of the neck, chest, abdomen and pelvis (4 scans).

Figure 3.

Average number of imaging studies acquired per patient according to year of diagnosis.

Discussion

In this favorable population of patients that achieved a complete metabolic response to frontline chemotherapy alone or combined modality therapy for limited stage HL, the relapse rate was low and the majority of relapses occurred within 2 years of completion of therapy. Routine surveillance imaging did successfully detect relapses, however 463 CT and PET-CT scans were performed for every relapse detected in the 2 years following treatment completion. Imaging surveillance with PET-CT imaging did not impact overall survival. Taken together, this data and previously published studies suggest limited utility for CT and PET-CT imaging surveillance in this favorable patient population^{10–12,22,31}.

In this study, patients treated for HL between 2003 and 2014 who received follow up care at our institution underwent a significant number of surveillance scans, however the number of scans performed decreased over time, reflecting the evolving recommendations for follow up imaging in this patient population. In the 2011 NCCN guidelines CT imaging was permitted every 6–12 months for the first 2–5 years after therapy²⁰. This is in contrast to the most recent NCCN guidelines that propose either clinical follow up or CT imaging at 6, 12 and 24 months after therapy⁸. Based on the current NCCN recommendations, CT imaging of the neck, chest, abdomen and pelvis may be ordered three times over the course of a 2 year follow up period which would result in 12 CT scans being ordered and a radiation dose of 66 mSV per patient. If every patient in our cohort had CT imaging performed according to current NCCN guidelines for the 2 year post-therapy period, then 2148 scans would have been conducted, corresponding to 269 scans performed for every relapse detected. Given that several studies suggest that early detection of relapsed HL does not confer improved responses to salvage therapy or a survival benefit, the current 6, 12 and 24 month strategy may be too aggressive^21,22.

There is no support for PET-CT surveillance imaging due to several studies that have demonstrated high levels of false positive findings on follow up PET-CT studies among HL patients^10–12,23. We did not find any differences in overall survival among the patients that had PET-CT imaging for surveillance during the 2 years after therapy compared to those that did not. All currently published guidelines agree that there is no role for PET-CT surveillance in HL patients in remission after frontline therapy^8,13,24. For patients that undergo combined modality therapy it is not clear whether PET-CT imaging should be repeated after RT is completed if a negative PET-CT scan was performed after systemic therapy. In the current study of the 135 patients that were treated with combined modality therapy, 126 patients had a negative PET-CT scan prior to RT. The majority of these patients (91%) had another PET-CT scan performed in the 2 year post-RT period. PET-CT alone did not detect any of the 3 relapses that occurred in this group of favorable patients. This suggests that repeating PET-CT imaging after RT when a pre-RT scan was negative may be of limited value.

Early stage HL patients that achieve a complete response on PET-CT imaging after frontline therapy have excellent outcomes^{15,16,25–27}. Even among more aggressive non-Hodgkin lymphoma subtypes that are associated with higher rates of relapse, surveillance imaging seems to have a limited role^7,28,29. Indeed in this group of HL patients treated at our institution, only 5% of patients experienced HL relapse, suggesting that aggressive surveillance imaging strategies are not justified. In our patient cohort, 75% of patients received consolidative RT, which has been shown to decrease rates of HL recurrence^25–27. Among patients treated with combined modality therapy, due to enhanced probability of cure, follow up imaging has even less value. Gandikota et al examined the need for follow up imaging among 78 patients with stage I/II HL treated at MSKCC with combined modality therapy³⁰. At a median follow up of 46 months, no relapses were observed and a total of 466 surveillance imaging studies were performed, corresponding to estimated excess cancer risk of 0.5%. Similarly Patel et al evaluated surveillance imaging among 78 patients treated with combined modality therapy for early stage HL³¹. In their study in which 32% of patients had bulky disease at diagnosis, 9 relapses were detected and 2440 follow up imaging studies were conducted. Only 3 relapses were detected by imaging. Taken together these studies coupled with our findings suggest that among early stage HL patients that fully respond to frontline treatment (especially combined modality therapy) surveillance imaging has limited utility.

For the nine relapses in our study, the median lymph node size was 2.9 cm, which is well below the detection level of a chest x-ray. This suggests that even follow up x-ray imaging may be of limited value as a follow up tool. Other alternative imaging strategies include magnetic resonance imaging (MRI) and ultrasonography (US), both of which have the advantage of no excess radiation exposure^32,33. Picardi et al performed a prospective study that randomized 300 advanced stage HL patients to either PET-CT or US and chest radiography³³. With 60 months median follow up, the relapse rate was 27% in this higher risk population and both follow up approaches were successful in identifying relapses. US/chest radiography demonstrated higher specificity and positive predictive value than PET-CT. The estimated cost per relapse was 10 fold higher with PET-CT and the exposure to radiation was 145 times higher with PET-CT (14.5 mSv versus 0.1 mSv). Ultimately though, with such low rates of relapse in the early stage HL population, even these imaging strategies may not be justified. One future strategy may be to utilize serum-based approaches such as circulating tumor DNA to identify patients at high risk for relapse that may benefit from a more intensive and costly surveillance strategy^34,35.

We found that 463 studies were performed per relapse in the first 2 years of follow up for the 179 patients in our study. This corresponds to estimated excess lifetime cancer risk of 0.6%. For patients in our study followed beyond 2 years this risk was estimated to be 1.1%. These predictions align with those previously reported in this patient population. In one study the mean increase in cancer risk was 0.4% and the maximum estimate was 1.2%¹⁸. Gandikota et al observed lifetime excess risk of 0.5% and maximum excess risk of 1.3%³⁰. It is notable that the majority of patients in these studies and our cohort received RT and the radiation dose from scans in comparison is low. However, it is still important to reduce additional exposure as much as reasonably achievable.

This study and the conclusions therein are limited by several factors beyond the retrospective nature of the study. The median follow up was adequate to assess HL outcomes however not long enough to assess the role of surveillance imaging in detecting second cancers. There may be a role for surveillance imaging in the late follow period to assess for treatment related toxicities. For the calculation of radiation exposure we utilized the higher end of radiation dose range, therefore radiation exposure may be overestimated in some patients. We also did not account for patient related factors that may influence decisions regarding follow up imaging frequency. Many patients experience anxiety after completion of therapy that is often alleviated by negative imaging studies. In addition, the majority of patients in our study cohort received consolidative RT. As treatment paradigms shift to decreased utilization of RT, increased relapse rates may justify routine surveillance imaging in certain patient populations. We also did not evaluate how many falsely positive surveillance studies occurred in our patient population which can be associated with costly workup and biopsies as well as patient anxiety. Finally, since the time period investigated in our study, improved imaging methods have been developed which have led to lower radiation dose utilization thus impacting lifetime excess cancer risk predictions.

Conclusions

We conclude that this work challenges the current guideline-specified routine surveillance imaging recommendations in HL. Risk stratified surveillance approaches may be considered for HL patients, in which limited stage patients that completely respond to frontline therapy with either systemic therapy alone or combined modality therapy, may be followed clinically without planned imaging surveillance.

Highlights.

• Surveillance imaging for Hodgkin lymphoma (HL) patients is controversial

• Among 179 limited stage HL patients 463 scans were performed to detect one relapse 2 years post therapy

• For early stage HL patients who completely respond to therapy imaging surveillance lacks clinical benefit