Immunotherapy targets regulatory pathways in T-cells to enhance antitumor immune responses and has dramatically changed the landscape of cancer treatment.1 Programmed cell death-1 (PD-1) inhibitors, such as nivolumab and pembrolizumab, and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitors, such as ipilimumab, are particularly well-researched targets. The use of these immune checkpoint inhibitors is expanding rapidly in several malignancies.2 At the same time, clinicians are facing new challenges related to the selection of patients who could most benefit from these agents as not all do. Thus far, no specific criteria have been established as an accurate predictive indicator for checkpoint inhibitors. The ideal duration for therapy is another important issue, given the high cost of immunotherapy and possibility of severe side effects. A strategy to maximize the potential therapeutic benefit, reduce toxicity, and manage cost is crucial for the field of immunotherapy.
18F-FDG PET is a widely used noninvasive method for the evaluation of tumors and their therapies. Several papers have recently reported the use of baseline and follow-up FDG-PET after starting immunotherapy in metastatic melanoma3 and Hodgkin lymphoma,4 primarily focused on tumor image parameters. Additionally, because the main effect of immunotherapy is the amplification of immune response, it is also likely worthwhile to evaluate the metabolism of lymphoid cell-rich organs, both for expected and excessive immune response. The latter can cause immune-related adverse events (IRAEs).5 Within IRAEs, there is a high frequency of hypophysitis with ipilimumab6 and thyroiditis with PD-1 antibodies.7 Because not all IRAEs present with clinical signs and symptoms, it is worthwhile to see if PET-detectable IRAE can provide additional information of clinical value. Moreover, there is discussion whether IRAEs could be a favorable prognostic marker for immunotherapy because IRAEs may reflect the flared immune activity needed for an antitumor effect. Several lines of evidence have indicated that IRAEs are associated with a higher response rate to immunotherapy,8 although this is still controversial. Based on this rationale, we hypothesized that IRAEs that are detectable by FDG-PET could provide an evaluation of the therapeutic effect of immunotherapy.
The goals of this study were to investigate whether the evaluation of FDG accumulation in tumors, lymphoid cell-rich organs, IRAE-related organs, and the development of IRAEs can predict the efficacy and outcome of immunotherapy, especially focusing on early time-point change of PET parameters, and whether PET-detectable IRAEs have any contribution for clinical management.
PATIENTS AND METHODS
Patients with advanced malignant disease (malignant melanoma [MM], malignant lymphoma [ML], and renal cell carcinoma [RCC]) who started immune checkpoint inhibitors between December 2013 and December 2016 were consecutively enrolled in this retrospective study. Patients were included if they had an 18F-FDG-PET/CT at baseline and at least one restaging scan. This study was approved by the institutional review board with HIPPA compliance at our facility, and informed consent was waived by institutional review board. We applied 2 categories to evaluate the efficacy of immunotherapy; best overall response in 1 year (BOR), and final outcome at the last follow-up; BOR was defined as the best tumor response during the 1-year follow-up period to see the use of an early time-point scan. The early time-point scan was defined as the first restaging scan. Tumor response was evaluated according to iRECIST9 for MM and RCC, and LYRIC10 for ML. In this study, iCR (interim complete response), iPR (interim partial response), iSD (interim stable disease), and iPD (interim progressive disease) used in iRECIST were described as CR, PR, SD, and PD as a matter of convenience. Patients with CR, PR, and SD in BOR were grouped as clinical benefit (CB) and those with PD in BOR as no-clinical benefit (No-CB). The final outcome was determined from a combination of data from the last follow-up PET scan and their clinical course and divided into either favorable or unfavorable outcomes. A favorable outcome included patients who continued on immunotherapy, did not have an exacerbation of disease after stopping immunotherapy, or patients with ML who had successful bone marrow transplantation after immunotherapy. Unfavorable outcomes included progression of disease after stopping immunotherapy, changing to chemotherapy, or death.
Patients were asked to fast for 6 hours before FDG injection. PET/CT images were obtained in 3-dimensional mode from the base-of-skull to mid-thigh for ML and RCC patients and from the vertex to the toes for MM patients. The activities of FDG were 9.83 ± 1.21 mCi. The time from injection to the start of the PET/CT scans was 69 ± 19 minutes for baseline and 65 ± 15 minutes for early time-point scan. The baseline and first restaging PET/CT scans were acquired on the Discovery 600, 690, or 710 (GE Healthcare, Chicago, IL), whereas during the follow-up phase, a fourth scanner, the GE MI was also used.
Maximum standardized uptake value (SUVmax) of viable tumors, whole-body metabolic tumor volume (MTV), whole-body total lesion glycolysis (TLG), and SUVmax of lymphoid cell-rich organs and IRAE-related organs were measured on each FDG-PET/CT scan using an AW workstation version 4.4 (GE Healthcare), with the consensus of 2 nuclear medicine physicians (T.N. and L.B.). Metabolic tumor volume was defined as the sum of whole viable tumor volume with an SUVmax cutoff of 2.5. Total lesion glycolysis was defined as sum of integration of mean SUV and MTV of each lesion. The SUVmax of lymphoid cell-rich organs and IRAE-related organs on the baseline and first restaging scans were also collected. Lymphoid cell-rich organs were defined as the spleen and bone marrow, and IRAE-related organs were defined as the pituitary and thyroid, which show a higher frequency of IRAE. Any lesions with heterogeneous or high physiological uptake from other surrounding tissue were avoided from region of interest. SUVmax of the IRAE-affected organs was also calculated. Symmetric uptake in lymph nodes was attributed to reactive lymphadenopathy and categorized as IRAE, with a confirmation on follow-up imaging. To see if early time-point changes of PET parameters are useful to estimate the efficacy of immunotherapy in 1 year, relationship between BOR and interval changes of SUVmax (Δ SUVmax), MTV (Δ MTV), TLG (Δ TLG), SUVmax of lymphoid cell-rich organs, and IRAE-related organs (Δ organ) between the baseline and first restaging scan was analyzed with Wilcoxon single rank sum test. The Mann–Whitney U test was used on the baseline parameters to ensure no statistical difference exists between CB and no-CB patients. Subsequently, the time course of SUVmax, MTV, and TLG of tumors was correlated with final outcome and patients who achieved CR with PET within 1 year were compared with final outcome using Fisher exact test. Finally, the temporal change in SUVmax of the IRAE-affected areas was measured if applicable. The number of days to reach peak of SUVmax were evaluated using the Steel-Dwass test. All statistical analyses were performed by T.N. using R version 3.4.2. and 2-sided P < 0.05 was considered significant.
Forty patients were included in total (male, n = 18; mean age, 57; MM, n = 21; ML, n = 11; RCC, n = 8) (Table 1). Three immune checkpoint inhibitors were used in this population (nivolumab, n = 19; pembrolizumab, n = 18; ipilimumab, n = 3). The mean duration of immunotherapy was 325 days (range, 42–1309 days), and the median follow-up period was 378 days (range, 97–1544 days). During the follow-up period, 33 patients suspended immunotherapy because of progression of disease (n = 10), side effects (n = 8), scheduled bone marrow transplantation (n = 8), or favorable response to therapy (n = 7). Five patients passed away during the follow-up period because of worsening of the disease (n = 3) or pneumonia (n = 2). CR, PR, SD, and PD in BOR were seen in 21, 7, 5, and 7 patients, respectively, and CB and no-CB were seen in 33 and 7 patients, respectively. Among 33 patients with CB, 5 patients had unfavorable outcome. Eventually, the favorable and unfavorable outcome was seen in 28 and 12 patients, respectively.
Relationship Between BOR and Early Time-Point Change of Tumor Parameters, Lymphoid Cell-Rich Tissues, and IRAE-Related Organs
Mean interval ± standard variance between the date of starting therapy and the first restaging scan was 91 ± 38 days. There was no statistical difference for baseline SUVmax, MTV, nor TLG between CB and no-CB patients (Table 2). Significant decrease in first restaging scans in SUVmax, MTV, and TLG was observed in CB patients, which was not seen in no-CB patients. Waterfall plots of ΔSUVmax, ΔMTV, and ΔTLG are shown in Figure 1. As for lymphoid cell-rich organs and IRAE-related organs, none of them showed significant results except for the spleen in no-CB patients, which showed a significant increase of SUVmax. All patients with increase above 1.5 in the thyroid (n = 5) and any decrease of SUVmax in the spleen (n = 18) showed CB (Fig. 2).
Time Course of PET/CT Parameters of Tumors
The time course of tumor parameters during the initial 24 months after starting immunotherapy is shown in Figure 3. All patients who achieved CR in the first 12 months (n = 22) showed a favorable outcome except for 1 patient who became refractory to therapy after 28 months. This was statistically associated with a favorable outcome (P = 0.002).
PET-Detectable IRAE and Its Relationship With Clinical Manifestations and Outcomes
Eleven patients developed IRAE which was detectable with PET/CT during immunotherapy, and six of them had multiple IRAEs (Table 3). Asymptomatic thyroiditis was incidentally found in 6 patients in PET. Two patients already showed findings of arthritis on the baseline scan, but these were clearly exacerbated after immunotherapy. Findings of colitis and pancreatitis were seen in 1 patient concurrently on PET/CT. Reactive lymphadenopathy was symmetrically seen in the mediastinum in 2 patients, in the bilateral inguinal area in 2 patients, and in the bilateral axillae in 1 patient. Steroid treatment was administered for 2 patients with cessation of immunotherapy after evaluation of PET. A spontaneous improvement was seen in 2 patients with thyroiditis and 2 patients with arthritis during follow-up PET studies. The other patients either continued to show IRAE findings during follow-up scans, or no further follow-up studies were acquired.
With respect to tumor response, 9 patients (82%) were confirmed as CR at the last follow-up scan. The time course of SUVmax in each affected lesion is shown in Figure 4. The mean date to reach a peak of SUVmax was 142 days in thyroiditis, 253 days in lymphadenopathy, 308 days in arthritis, and 283 days in both colitis and pancreatitis. Peak SUVmax of thyroiditis was observed significantly earlier than that of arthritis. Representative cases are shown in Figures 5 and 6.
The goal of this study, in patients starting immunotherapy, was to understand the following: 1) whether an early time-point change of SUVmax, MTV, and TLG of tumoral lesions, and SUVmax of lymphoid cell-rich organs, and IRAE-related organs on an FDG PET scan, can predict the clinical benefit of immunotherapy in 1 year, 2) whether PET parameters of tumoral lesions such as SUVmax, MTV, and TLG are correlated with final clinical outcomes, and 3) whether PET-detectable IRAEs are of short-term clinical significance and/or predictive of final outcomes. If so, PET can be used for improved patient selection for this potentially beneficial but also expensive and potentially toxic therapy. Our results show that patients who had clinical benefit in 1 year with immunotherapy showed a significant decrease of FDG uptake with respect to SUVmax, MTV, and TLG and showed either an increase of SUVmax in the thyroid gland of more than 1.5 or a decrease of SUVmax in the spleen. Patients who had CR with FDG-PET within 1 year of starting immunotherapy were significantly associated with a favorable outcome. PET-detectable IRAEs were seen in 11 patients, and 2 of them changed therapeutic strategies, whereas 9 of them showed CR as their final outcome. Among IRAE, thyroiditis was seen significantly earlier than arthritis.
There are several articles which have investigated the relationship between FDG tumor parameters and effectiveness of immunotherapy. Tumor PET parameters were reported as accurate biomarkers for the efficacy of immunotherapy,3,4 whereas there are also a few negative reports that PET parameters were unreliable in predicting treatment response.11,12 In our study, interval change of tumor parameters at an early time point was significantly decreased in the CB population, but none of these values were significant. Considering this, tumor PET parameters are assumed to be somewhat reliable but still remain an imperfect prognostic indicator as FDG accumulates in both tumor and sites of inflammation where immune cells are stimulated.3
Aside from the tumors, we evaluated whether FDG accumulation in lymphoid cell-rich organs, including the spleen and bone marrow may provide additional information. Dercle et al showed that a significant increase of SUVmax in the spleen at 3 months was seen in responders to PD-1 inhibitors in Hodgkin lymphoma.4 On the other hand, our results show that a significant increase of SUVmax in the spleen was seen in patients who did not have clinical benefit from immunotherapy. One of the reasons why our no-CB population showed elevated SUVmax in the spleen might be due to the fact that average SUVmax at baseline in CB patients was higher than that of no-CB patients, although the difference was not statistically significant. As such, the significance of FDG uptake in the spleen remains controversial, and validation with additional studies is needed.
Immune-related adverse event is a frequently seen event in immunotherapy, although its precise pathophysiology remains unclear.13 In our study, 7 patients of 11 had PET-detectable IRAE without any signs or symptoms. PET-detectable IRAEs can therefore sometimes be asymptomatic and clinically missed. Moreover, a few cases required a change in management. PET-detectable IRAEs are important to alert the clinician for careful monitoring and possibly changing management, such as discontinuing immunotherapy or initiating treatment for IRAEs.
We also evaluated whether PET-detectable IRAE could be a favorable prognostic marker. Patients with IRAEs showing an increase in SUVmax of more than 1.5 in the thyroid gland on the early time-point PET had ultimate clinical benefit from the therapy. Diffuse thyroidal FDG uptake or autoimmune thyroiditis was previously reported to predict favorable outcome in diffuse large B cell lymphoma with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP)14 and in advanced breast cancer with chemotherapy.15 As such, thyroiditis may reflect the activated status of the immune response against cancer as therapy-induced autoimmunity. We also evaluated IRAEs with respect to their time course and correlation with final clinical outcomes. Approximately 82% of patients with IRAEs had a complete response on the final restaging scan. The average time to reach the peak SUV in thyroiditis was shortest among PET-detectable IRAE, and this was significantly shorter than that of arthritis. This was comparatively consistent with previous article, which showed that thyroiditis was mostly observed within the first weeks after the initiation of immunotherapy.7 Considering this, elevated diffuse FDG uptake in thyroid at an early time-point PET may be a favorable indicator in immunotherapy.
This study has several limitations including the small number of patients and heterogeneous malignancies and therapies. This heterogeneity was possibly the cause of the fact that IRAEs were not observed in either RCC patients or patients with anti CTLA-4 antibody. Another limitation would be that no pathological confirmation was eligible from tissues suspected to have IRAEs as this retrospective study was not designed in that manner. This study had a relatively long period of follow-up (average, 19 months), but an even longer follow-up period may have provided more accurate results. To overcome these limitations, future studies with a larger cohort are needed to validate our results.
Early time-point FDG-PET/CT evaluation after starting immunotherapy shows that PET parameters associated with clinical benefit in 1 year included a significant decrease in tumoral FDG uptake. Furthermore, patients who had CR by PET within the first year of starting therapy ultimately had a favorable final clinical outcome. PET was also considered as an important modality to evaluate IRAE. PET-detectable IRAE indicated a favorable outcome, and in particular, thyroiditis was seen earlier than other IRAE and could provide an early indicator of the efficacy of immunotherapy, although validation with a larger cohort is needed.
1. Sharma P, Allison JP. The future of immune checkpoint therapy. Science
2. Friedman CF, Proverbs-Singh TA, Postow MA. Treatment of the immune-related adverse effects of immune checkpoint inhibitors
: a review. JAMA Oncol
3. Anwar H, Sachpekidis C, Winkler J, et al. Absolute number of new lesions on 18
F-FDG PET/CT is more predictive of clinical response than SUV changes in metastatic melanoma patients receiving ipilimumab. Eur J Nucl Med Mol Imaging
4. Dercle L, Seban RD, Lazarovici J, et al. 18
F-FDG PET and CT scans detect new imaging patterns of response and progression in patients with Hodgkin lymphoma treated by anti-programmed death 1 immune checkpoint inhibitor. J Nucl Med
5. Postow MA, Sidlow R, Hellmann MD. Immune-related adverse events associated with immune checkpoint blockade. N Engl J Med
6. Faje A. Immunotherapy and hypophysitis: clinical presentation, treatment, and biologic insights. Pituitary
7. Yamauchi I, Sakane Y, Fukuda Y, et al. Clinical features of nivolumab-induced thyroiditis: a case series study. Thyroid
8. Bronstein Y, Ng CS, Hwu P, et al. Radiologic manifestations of immune-related adverse events in patients with metastatic melanoma undergoing anti-CTLA-4 antibody therapy. Am J Roentgenol
9. Seymour L, Bogaerts J, Perrone A, et al. iRECIST: guidelines for response criteria for use in trials testing immunotherapeutics. Lancet Oncol
10. Cheson BD, Ansell S, Schwartz L, et al. Refinement of the Lugano Classification lymphoma response criteria in the era of immunomodulatory therapy. Blood
11. Sachpekidis C, Larribere L, Pan L, et al. Predictive value of early 18
F-FDG PET/CT studies for treatment response evaluation to ipilimumab in metastatic melanoma: preliminary results of an ongoing study. Eur J Nucl Med Mol Imaging
12. Ribas A, Benz MR, Allen-Auerbach MS, et al. Imaging of CTLA4 blockade-induced cell replication with (18)F-FLT PET in patients with advanced melanoma treated with tremelimumab. J Nucl Med
13. Freeman-Keller M, Kim Y, Cronin H, et al. Nivolumab in resected and unresectable metastatic melanoma: characteristics of immune-related adverse events and association with outcomes. Clin Cancer Res
14. Song MK, Chung JS, Kim SJ, et al. Diffuse thyroid 18
F-FDG uptake after R-CHOP therapy predicts favorable outcome
in patients with DLBCL. Ann Hematol
15. Fiore E, Giustarini E, Mammoli C, et al. Favorable predictive value of thyroid autoimmunity in high aggressive breast cancer. J Endocrinol Invest