1. IntroductionThe emergence of immune checkpoint inhibitors has resulted in significant progression in the treatment of advanced non-small cell lung cancer (NSCLC) [1,2,3]. Administered as monotherapy or in combination with chemotherapy, PD-1/PD-L1 inhibitors result in improved overall survival (OS) of stage IV NSCLC among untreated and pretreated patients [1,3,4]. However, the radiological evaluation of PD-1/PD-L1 inhibitor-treated NSCLC is challenging due to the occurrence of atypical response (AR), manifesting as simultaneous regression in some tumors and progression in others [5]. The development of AR complicates the evaluation of treatment efficacy and decisions regarding subsequent treatment strategies. Although several versions of imaging-response standards related to the PD1/PD-L1 inhibitor have been explored to shrink the impact of AR patterns with regard to tumor-response evaluation [6,7], real-world clinical descriptions of this immune-related response pattern and treatment strategy in NSCLC remain scarce. Moreover, the prognostic significance of AR is not yet fully understood. Studies reported that the prognosis of patients with AR was superior to those of patients with typical progression but inferior to those with a typical response (TR) [5,8,9,10]. However, these studies had limited sample sizes and were generally insufficient to capture the full picture of AR in advanced NSCLC. For example, the radiographic patterns and clinical management of patients with AR have seldom been investigated.Local consolidative therapy is an interesting topic for the treatment of oligo-progressive sites, and it has been increasingly provided in clinical practice. A wealth of research has demonstrated the essential role of local cancer treatments in improving survival among NSCLC patients with oligo-progressive or oligometastic disease [11,12,13]. Radiotherapy has immunomodulatory qualities capable of augmenting antitumor immune responses, making the integration of radiotherapy with immunotherapy a new therapeutic option in advanced NSCLC [14,15]. However, few studies have commented on treatment recommendations for patients with AR in NSCLC. Here, a retrospective, multicenter, collaborative study was performed to investigate the clinical characteristics, treatment strategy, and survival results of AR in metastatic NSCLC patients treated with PD1/PD-L1 inhibitor. 4. Discussion

In the current study, a multicenter retrospective analysis was conducted to investigate the clinical features, prognostic significance, and treatment of patients with stage IV NSCLC with AR receiving immune checkpoint inhibitors. The outcomes of PD1/PD-L1 inhibitor in patients with AR were comparable to those in patients with TR. Moreover, proper local therapy targeting paradoxical progressive lesions may provide survival benefits for patients with AR, which warrants further validation.

The clinical features of AR are not fully understood, and this study provides significant information about the frequency and spatiotemporal patterns of AR in patients with metastatic NSCLC treated with PD1/PD-L1 inhibitor. The frequency of AR was reported to be from 5.0% to 13.0% in PD-1/PD-L1 inhibitor-treated metastatic NSCLC [5,10,20], which was consistent with the current study (6.1%). Previous studies [5,9] reported a median interval to AR of within 3 months, which was similar to the finding of the present study. Hence, the paradoxical progressive disease that developed during the first 2–3 months needed to be interpreted with caution because a non-negligible percentage of patients may actually have AR. In addition, all lesions during the initial radiological examination were evaluated, and patients with more than three metastatic organs or sites were found to be more likely to develop AR. The spatial distribution of progressive lesions in patients with AR, in terms of sites and number, was quite similar to that reported in those with acquired resistance to PD1/PD-L1 inhibitor [5,21,22,23,24,25]. In both clinical scenarios, the common sites of progressive disease often occurred in the lymph nodes and lung, and the most common progressive lesion mainly occurred in the originally existing sites; meanwhile, the majority of patients had no more than three progressive lesions, indicating potential shared mechanisms regarding intra-tumoral heterogeneity and reductant immune evasion pathways in metastatic NSCLC [22,26,27,28]. Accordingly, a previous study reported that acquired resistance to the PD1/PD-L1 inhibitor usually occurred in one or two sites of disease when local treatment and the continuous PD1/PD-L1 inhibitor led to prolonged benefits [22], in line with the results of the current research.However, the mechanism underlying AR remains unclear. The heterogeneity and differences of intra-tumoral between metastatic sites may contribute to fluctuations in clinical effects (i.e., some tumor lesions shrinkage while others growth). Moreover, it takes some time to activate the immune system, the start of clinical effects may be delayed, and the interaction between the immune system and the tumor may be a long-term process that could possibly contribute to some sites shrinking while other sites are grow or new lesions emerge [20]. PD-L1 expression, as an important biomarker for immune checkpoint inhibitors and an efficient and crucial checkpoint of acquired immune resistance in NSCLC, was found to be different among tumor specimens of different sites in patients with NSCLC [29,30,31], highlighting the magnitude of intra-tumoral immune heterogeneity. A notable detail is that tumor microenvironments, such as T cell recruitment and infiltration, may influence the outcome of PD-1/PD-L1 inhibitors and differ across metastatic organs in patients with tumor [32]. Meanwhile, patients with AR often experienced concomitant increases in some lesions and a decrease in other lesions, indicating that patients with AR could benefit from the PD1/PD-L1 inhibitor and continued PD-1/PD-L1 inhibitor combination treatment beyond the first disease progression on RECIST version 1.1 compared with patients who stopped the PD1/PD-L1 inhibitor and changed anticancer therapy [23,33,34,35]. Therefore, the continuation of the PD1/PD-L1 inhibitor beyond progression may be appropriate for patients with AR in advanced NSCLC.The value of local therapy for stage IV NSCLC patients in an AR cohort treated with PD-1/PD-L1 inhibitors has seldom been reported. A retrospective study reported that the duration of PD1/PD-L1 inhibitors could be amplified by local therapy, such as radiotherapy, to solitary or oligo-progressive sites in atypical response, especially among those harboring a biologically more indolent cancer type, such as renal cancer [24]. Moreover, surgery or radiotherapy are becoming an integral component in the treatment of oligo-progressive disease in NSCLC [36]. Local consolidative therapy with or without maintenance therapy for patients with three or fewer metastases from NSCLC that did not progress after initial systemic therapy improved progression-free survival and overall survival compared with maintenance therapy alone [13,37]. In the present study, most progressive tumor lesions developed from the originally existing tumor sites in patients with AR. Proper local therapy for progressive lesions in patients with AR could prolong PFS in NSCLC and thus extend the PD1/PD-L1 inhibitor treatment duration, which is generally related to less toxicity and better quality of life than conventional chemotherapy [1]. The current findings regarding the clinical value of local therapy within the PD1/PD-L1 inhibitor window in patients with AR were comparable to those reported in patients with metastatic NSCLC treated with the PD1/PD-L1 inhibitor who developed acquired resistance [22,23,38]. Local therapy, such as radiotherapy, targeting oligo-progressive tumor lesions in patients with AR or acquired resistance may improve patient survival as it could effectively eradicate tumor lesions compromising PD1/PD-L1-inhibitor-resistant subclones and enhance the systemic antitumor immune response through synergistic effects between radiotherapy and immunotherapy [38,39,40]. However, the current work may be the first to report the potential clinical benefit of local therapy for progressive targeted sites or new lesions in patients with AR in advanced NSCLC, thereby warranting future investigation.

This study has some limitations. First, selection bias and unrecognized confounding factors could exist between subgroups, in which survival outcomes were compared. Although potential bias was reduced as much as possibly by using Cox analyses and PSM, imbalanced baseline characteristics could be inevitable. Thus, the findings in the present study need to be validated by prospective randomized trials. Second, as a retrospective study, collecting detailed information about treatment toxicities was difficult, and thus the safety profile of adding local therapy needs to be investigated in further prospective studies.