Here, we conducted a retrospective analysis of 131 unselected stage IV cancer patients with 23 different cancer types who were treated with T-cell-stimulating modalities, hyperthermia, IL-2, and ipilimumab plus nivolumab (0.5 mg/kg and 0.3 mg/kg, respectively). Twenty-four percent of our patients had to receive antibiotics. Antibiotic treatments administered within 30 days from commencement of ICI therapy has been known to be associated with significantly worse overall survival in patients with nonsmall cell lung cancer (2.5 vs. 26 months), melanoma (3.9 vs. 14 months), and other tumor types (1.1 vs. 11) who received antibiotic treatment vs. those who did not. In addition, patients had a higher risk of disease refractory to treatment .
During a follow-up period of up to 5 years, the ORR and OR were 31.3% and 49.62%, respectively, while the OS at 12 months was 65.9% (Table 2). irAEs of WHO grade 1, 2, 3, and 4 were observed in 23.66%, 16.03%, 6.11%, and 2.29% of patients, respectively (Table 3). In other words, less than half (48.09%) of the patients experienced irAEs of any grade, while only 8.4% had grade 3 or 4 irAEs, but no treatment-related death occurred. Therefore, our protocol does not require selection of stage IV cancer patients who are more likely to gain benefit from ICIs. Our safety results compare favorably to that of the study by Callahan et al. but also to the meta-analysis of Xu et al., in both of which registered doses of concurrent nivolumab and ipilimumab treatment were administered. Any-grade irAEs occurred in 96.8% of patients in Callahan’s study, while 58.5% of patients had grade 3 and 4 irAEs leading to discontinuation in 24.5% of patients, and one treatment-related death was also reported . Similarly, grade 3–4 irAEs occurred in 39.9% of patients in the meta-analysis of Xu et al., while 2.0% treatment-related death was recorded .
The combination of ipilimumab and nivolumab has first been established in melanoma  and renal cell cancer  but also described for nonsmall cell lung cancer [33, 34]. In these patients, it is well known that high mutational burden favors better response rates . Although immunotherapy has become one of the greatest advances in oncology over the last century, the application for the treatment of breast cancer remains an area of investigation . Consistent with this, we found 1365 papers with the key words < cancer, ipilimumab, nivolumab > but only 18 with the key words < breast cancer, ipilimumab, nivolumab > (PubMed search as of June 2020). Based on our data presented in this paper, such neglect of breast cancer seems to be unjustified. We estimated the hazard ratio (HR) for the difference between all the 131 cancers and 42 breast cancers using a Cox proportional hazard model and found no significant differences (Fig. 4a, b). 95% confidence intervals for HRs provide ranges of plausible values for the true HRs of our data.
Importantly, the overwhelming majority (> 98%) of the evaluated patients in this study had low PD-L1 expression as confirmed by immunohistochemistry (determined as ≤ 1%). Low tumor mutational burden (TMBlow) and low microsatellite instability (MSIlow) was determined only in a small subgroup of patients. However, TMBlow and MSIlow can be expected in the majority based on published evidence . In contrast, published meta-analyses included patients with high PD-L1 expression, TMBhigh, and MSIhigh [30, 38]. Therefore, in contrast to the published meta-analysis, our patient group had the following negative pre-selection factors:
Antibiotic use in 24%;
Low PD-1/PD-L1 expression;
Only stage IV patients, with 35.1% liver metastasis, a specifically bad prognosis;
Only 35% ECOG 0;
Heavily pretreated patients.
Control of minimal residual cancer by exploiting autoimmunity induced by low-dose ipilimumab was proposed earlier . For safety reasons, we administered the lowest doses of nivolumab (0.5 mg/kg) and ipilimumab (0.3 mg/kg), which induced grade 3 or 4 irAEs in only 8.4% of patients (Table 1.). The low dose ICI protocol was justified by Sen et al., who demonstrated that despite a dose-dependent increase in irAEs, no improvement in PFS, OS, or disease control rate (DCR) were identified with escalating doses of ICIs . The authors concluded that lower doses may reduce toxicity and cost without compromising disease control or survival. In fact, due to the rapidity of development, competition, and race for FDA approval, the optimal dosing and schedule of ICIs are still not fully defined and continue to be under study .
Weber predicted 10 years ago that abrogation of the CTLA-4 function results in immune stimulation, tolerance breakdown and eventually tumor eradication . This prediction has recently been formally confirmed by Eggermont et al. in a study including 1019 adults with stage III melanoma, in which patients were randomly assigned on a 1:1 ratio to receive treatment with pembrolizumab therapy or placebo . The occurrence of an irAE was indeed associated with a significantly longer recurrence-free survival (RFS) in the pembrolizumab arm.
ICIs are more likely to halt tumor growth in patients with a higher TMB than in those with a lower one [43, 44]. According to the consensus of experts, mutational load in cancer cells (e.g., lung cancer patients with former nicotine abuse) may generate novel antigens that are not subject to immune tolerance and allow for an adaptive immune response by the host. We proposed an alternative interpretation for the induction of immune response. Tumor cells with newly expressed neoantigens are no longer recognized as “self” because they are transformed into “non-self” such that they become targets for the patient’s own immune system. Owing to the newly expressed neoantigens the unresponsiveness/tolerance that existed between the patient’s immune system and cancer cells was abolished. This in turn, resulted in the development of an auto-GVHD with secondary therapeutic benefits, in analogy with the GVM effects following allogeneic stem cell transplantation . Although a limited transformation is too weak in itself to instigate a tumor eradicating T-cell attack, with immune checkpoint blockade T cells are more effective against “altered self” resulting in better OS . Not unexpectedly, a significant positive correlation was found between the reporting odds ratio (ROR) of reporting an irAE during anti-CTLA-4, anti-PD-1, and anti-CTLA-4/anti-PD-1 immune combination therapies and the corresponding TMB in 7677 patients across 19 cancer types . Consistent with this, Berner et al. demonstrated in NSCLC that T cells recognize and target shared tumor and skin antigens during ICI therapy resulting in autoimmune-mediated skin toxicity and tumor regression .
Fever therapy and checkpoint inhibitors
Combining thermal therapy with immune therapy should greatly increase the response rate of immune therapy as all types of hyperthermia appear to increase the numbers of effector lymphocytes [48, 49]. Inducing one week of daily cyclic high fever response during individually dose adapted IL-2 we are in the footsteps of William B. Coley, the father of cancer immunotherapy who administered an FDA-approved fever inducing bacterial inoculate for the treatment of soft tissue sarcomas . It is worth to recall that in 1891 Coley was convinced that stimulating the immune system by a severe infection (such as erysipelas), which was associated with high fever, would have the side effect of shrinking the malignant tumor . By the end of his career, Coley had written over 150 papers and treated almost 1000 cases and noticed that in 500 of these there was near-complete regression .
Hyperthermia is almost always used in combination with other forms of cancer therapy [49, 53]. While many studies have shown a significant reduction in tumor size, just a few could demonstrate a moderately increased survival in patients receiving the combined treatments (e.g., ). Our response rates in stage IV patients with unfavorable MSIlow, PD-L1 < 1%, TMBlow, 26% of which received antibiotics are promising. Since the proposed protocol consists only of approved drugs and treatments, our hypothesis that low‐dose ICI‐induced autoimmune T cells are powerful therapeutic tools can be confirmed or refuted in controlled prospective clinical trials.
A retrospective analysis of case series cannot replace a controlled prospective clinical trial, the gold standard for the evaluation of new treatments. Furthermore, this study was underpowered to detect differences between survival rates in cancer subgroups. Notwithstanding, case reports and series represent relevant study designs, which can be highly influential in furthering medical knowledge despite of their limitations when a question of importance cannot be addressed by other methods because of ethical or logistical constraints [54, 55].