The development of immune-checkpoint inhibitors (ICI), i.e., anti-CTLA-4 and anti-PD-1/PD-L1, had a major impact on the management of solid tumors, particularly for metastatic melanoma and lung cancer, leading to a significant improvement in overall survival. Therapy response rate ranges between 15 and 50% in various series [1], thus the early identification of non-responders is crucial to allow a prompt switch to a following line of treatment, also to reduce costs and limit immune-related adverse events (irAEs). irAEs can occur in up to 40% of patients treated with single-agent immunotherapy but that can even be higher with ICI combinations [2]. Although solely dimensional criteria like RECIST were prevalently used in registrative studies of ICI, their peculiar indirect mechanism of action different from that of conventional chemotherapy and target therapy, led to the advent of atypical imaging response patterns complicating the oncological clinical decision-making process [3]. Among them, the so-called “pseudoprogression”, has created many issues in the assessment of response to treatment, particularly during the first weeks of therapy and in absence of clinical signs and/or symptoms of disease progression. This phenomenon, which consists of a transient increase in tumor volume and metabolism sometimes associated with the appearance of new lesions, has been attributed to several mechanisms including delayed response to treatment, infiltration of immune cells into the tumor and edema. Another atypical response pattern for immunotherapy is the “hyperprogressive disease” (HPD), consisting of a sudden acceleration of tumor growth, often associated with rapid clinical deterioration. Incidence of HPD is highly variable ranging between 4 and 30% among studies and it seems to be associated with a high pre-treatment metabolic tumor burden [4].

For the above reasons, efforts have been made to find response criteria specifically adapted for immunotherapy for both CT (irRC, irRECIST, iRECIST, imRECIST) [5] and PET/CT, namely PET Response Evaluation Criteria for Immunotherapy (PERCIMT) [6], PET/CT Criteria for early prediction of Response to Immune checkpoint inhibitor therapy (PECRIT) [7], Immunotherapy-modified PERCIST (imPERCIST) [8] and immune PET Response Criteria in Solid Tumors (iPERCIST) [9] (Table 1). These criteria were developed based on melanoma (PERCIMT, imPERCIST and PECRIT) and lung cancer (iPERCIST), the two solid tumors in which more experience has been reported on the clinical use of PET in immunotherapy. These new criteria are based on two main principles linked to the aforementioned nature of pseudoprogression:—it is advisable to reconfirm progressive disease in a short-term study (iPERCIST) [9];—the appearance of new lesions cannot be considered per se as disease progression (PERCIMT, imPERCIST) [6, 8]. PECRIT, on the other hand, is a mixed criterion which combines morphological (RECIST) and metabolic criteria (PERCIST) [7]. A recent joint practical/procedural guideline by EANM/SNMMI/ANZSM [10] clearly summarizes the above topics, also addressing others aspects related to immunotherapy such as the appearance of IRAEs and signs of activation of the immune system, whose prognostic significance is still controversial. Indeed, the increase of spleen FDG uptake during ICI was originally identified as a favourable predictor for immune activation, however, more recent studies have reported that this sign is an unfavourable finding, likely associated with cancer inflammation and tumor burden [11]. Many of the studies on 18F-FDG PET/CT response assessment in immunotherapy deal with metastatic melanoma treated with ipilimumab, the first approved anti-CTLA-4 antibody in 2011. Ipilimumab ushered in the era of new immunotherapics aimed at disrupting immunological tolerance. Its prevalent action on early T cell activation phase in central lymphoid tissue rather than on antigen experienced T cell in the tumoral microenvironment, like anti-PD1/anti-PD-L1 antibodies, leads to slower triggering of the immunological response with clinical responses rarely seen before at least 12 weeks from the start of treatment [12]. In fact, it is demonstrated that the rate of pseudoprogression is higher for ipilimumab than for anti-PD1 when a chemotherapy-like schedule of evaluation of response is used (every 8–12 weeks). Therefore, it seems that the timing of the first PET/CT re-evaluation during immunotherapy (which must necessarily be longer for Ipilimumab) rather than the response criterion to be adopted, is the key factor for a correct interpretation of the scan: the finding of a Progressive Metabolic Disease (PMD) at a PET/CT scan performed 12 weeks after the start of immunotherapy is more likely to be a true progression in patients treated with anti-PD1, while the possibility of a pseudoprogression should be considered in the case of treatment with ipilimumab. Furthermore, another concept that is emerging from some published studies is that the predictive value of immunotherapy PET/CT response criteria is not superior to conventional ones (PERCIST, EORTC) [13,14,15]. Thus, it can be assumed that the standard criteria can be conveniently used in clinical practice with some precautions. First, regardless of whether the PERCIST or EORTC criteria are adopted, it is preferable to select many target lesions (up to 5 lesions, max 2 lesions per organ), to take into account possible dissociated responses (approximately 10% of cases) which have been shown to have a better prognosis [16]. Second, in the event of progressive metabolic disease on PET imaging not associated with clinical worsening, additional details should be provided for the clinical decision support to continue or stop the treatment as follows: what is the extent of the increase in lesions uptake? If new lesions have appeared, what is their impact on the total tumor burden? Is there evidence of a dissociated response? Is there an oligoprogression?

Table 1 Overview of conventional and immune-specific response criteria with 18F-FDG PET/CT
Full size table

Another important issue is the prognostic significance of Stable Metabolic Disease (SMD), possibly representing a heterogeneous group of patients including both responders and non-responders, as demonstrated in a cohort of lung cancer patients treated with nivolumab, in which the survival curves for SMD overlapped those for PMD [14]. As an attempt to identify the subgroup of patients with SMD who are not responding to the immunotherapy treatment, it is advisable to check for any discrepancies between the variation of the SUVmax and the active metabolic volumes of the lesions. Indeed, it has been demonstrated that a subgroup of metastatic melanoma patients who did not respond to treatment with ipilimumab was shown to have increased metabolic volume of lesions despite stability or even decreased SUVmax values [13]. A final aspect to be stressed is the actual role of PET/CT in the clinical decision to discontinue treatment in responding patients to anti-PD1 (± ipilimumab)/anti-PDL-1 that, as per technical data sheet and in absence of comparative prospective clinical trials, should be continued until disease progression or unacceptable toxicities. Several retrospective analyses in melanoma have shown that the depth of response (CR vs PR vs SD) correlates with the chance of long-term maintenance of the response despite ICI interruption. This acquisition together with the prognostic impact of Complete Metabolic Response (CMR) on survival led the ESMO recommendations to incorporate the possibility to offer the discontinuation of immunotherapy in patients with metastatic melanoma in CMR once provided they have undergone treatment for at least 6 months, even in the presence of SD or PR at the TC [17].

All these concepts are summarized in Fig. 1, which illustrates a proposed diagnostic flowchart for the management of patients. Considering the well-known limitations of diagnostic accuracy of 18F-FDG PET/CT in the study of brain metastasis, patients should also be monitored by brain CT or MRI.

Fig. 1
figure 1

A flow chart for 18F-FDG PET/CT imaging in the management of patients treated with immunotherapy

Full size image

In conclusion, data on the use of 18F-FDG PET/CT in the evaluation of the immunotherapy response available so far are promising, however, further prospective studies are still needed to optimize the potential of metabolic imaging and to better understand its prognostic role.