Assessment of anti-PD-(L)1 for patients with coexisting malignant tumor and tuberculosis classified by active, latent, and obsolete stage

Tuberculosis and cancer represent two major challenges in health care worldwide, especially in China, where the burden of tuberculosis remains the third highest globally, and the coexistence of tuberculosis and cancer is not uncommon [1‐3]. The relationship between tuberculosis and cancer is complex [4]. Multiple studies have demonstrated that cancer is a risk factor for tuberculosis activation and tuberculosis itself also promotes cancer, especially lung cancer [4‐9].

The safety and efficacy of anti-cancer therapy in patients with coexisting malignancy and tuberculosis treated with concurrent anti-tuberculosis had been investigated in previous studies [10‐12]. Recently, we also confirmed the clinical benefits of the concurrent anti-tuberculosis and chemotherapy in lung cancer patients with co-existent tuberculosis [13]. Compared to conventional anti-cancer regimens, immune checkpoint inhibitors (ICIs) have proven to offer superior survival benefits and fewer adverse events [14], thus making them an emerging therapeutic option for cancer patients and even first-line treatment for patients with some advanced-stage malignancies. However, patients with autoimmune disease or chronic infectious disease including tuberculosis were routinely excluded from trials of checkpoint immunotherapies, as ICIs may cause host immune imbalance and develop immune-related adverse events (irAEs) [15].

A recent study had shown that ICIs could be a treatment option for patients with lung cancer and a history of tuberculosis [16]. However, whether ICIs could be used for patients with coexisting malignancy and tuberculosis, especially for those with active tuberculosis treated with concurrent anti-tuberculosis, remains unknown. Moreover, programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) pathways had been proven to be crucial in controlling excessive inflammation in tuberculosis, while deficiency in PD-1 would lead to deterioration of tuberculosis in animal models [17‐19]. Therefore, whether anti-PD-(L)1 would induce treatment failure in anti-tuberculosis regimen or even result in tuberculosis re-activation remains to be an unresolved concern for clinicians.

In this study, we assessed the efficacy and safety of anti-PD-(L)1 treatment in patients with coexisting malignancy and tuberculosis and investigated the efficacy of anti-tuberculosis therapy in those with active tuberculosis.

To our knowledge, this is the first real-world cohort study with the largest sample size to evaluate the efficacy and safety of anti-PD-(L)1 treatment in patients with coexisting malignancy and tuberculosis, as well as the efficacy of anti-tuberculosis treatment in those with active tuberculosis. Previous studies have reported individual cases, but there was no systematic cohort study [26‐28]. This study suggests that these patients could benefit from anti-PD-(L)1 immunotherapy while achieving unimpaired anti-tuberculosis therapy. The results of this study provide reasonable clinical evidence for the treatment of patients with coexisting malignant tumor and tuberculosis, and the concurrent anti-PD-(L)1 and anti-tuberculosis therapy in patients with active tuberculosis.

Previous studies have identified specific immune activation and immune microenvironment alteration during tuberculosis infection [29, 30]; however, whether tuberculosis affected the tumor microenvironment and anti-tumor therapy was unknown. We furtherly analyzed the immune microenvironment in the resected tumor specimens from lung cancer patients with active or obsolete tuberculosis. The results revealed that patients with active tuberculosis had higher rate of high expression of PD-L1 and CD8+ lymphocyte infiltration compared to those with obsolete tuberculosis or single lung cancer (Additional file 2: Fig. S1). These results, to some extent, supported that patients with coexisting malignant tumors and active tuberculosis may be the potential candidates who will benefit from PD-1 blockade immunotherapy due to their inflammatory microenvironment.

The implication of PD-1/PD-L1 pathway has been confirmed in the pathophysiology of tuberculosis in preclinical studies [17‐19, 31]. PD-1-deficient mice exhibited significant sensitivity to M.tuberculosis infection and survival reduction [19]. In addition, the infected PD-1-deficient mice developed severe necrotic pneumonia with marked elevation of serum proinflammatory cytokines [17]. These results show that the PD-1/PD-L1 pathway is involved in the occurrence and development of tuberculosis. Furthermore, both acute and reactivated tuberculosis have been described in patients undergoing treatment with anti-PD-1 in previous studies [28, 32, 33]. The view that tuberculosis could develop in cancer patients receiving immunotherapy has been represented in some studies [34]. However, the risk of reactivation of latent M tuberculosis or primary tuberculosis infection during ICI therapy is still unknown and clinicians are alerted to the development of active tuberculosis during immunotherapy. Thus, we aimed to address this clinical issue. In the latent tuberculosis group, none of these patients developed active tuberculosis after anti-PD-(L)1 therapy, which was in line with another retrospective study in which none of the anti-PD-1 regimen treated cancer patients with positive IGRA testing developed active tuberculosis [17]. In the active tuberculosis group, the combination of anti-PD-(L)1 and anti-tuberculosis treatment was safe but there were 2 patients with tuberculosis relapse; In the obsolete group, there was 1 patient with tuberculosis relapse. For these patients with tuberculosis relapse, it might be because of the reduced immunity in the condition of suffering from malignancy or the irregular anti-tuberculosis therapy due to the anti-PD-(L)1 treatment. Notably, 6 patients in our cohort initiated anti-tuberculosis therapy after anti-PD-(L)1 therapy (Fig 3). To explore whether they developed tuberculosis during anti-PD-(L)1 therapy, we reviewed their imaging data and found that suspicious tuberculosis lesions had already existed before anti-PD-(L)1 therapy and these lesions enlarged after anti-PD-(L)1 therapy but reduced after anti-tuberculosis treatment, resulting in a pseudo progression (Additional file 2: Fig. S2). These results indicated that the tuberculosis of these patients might be not caused or reactivated by anti-PD-(L)1 agents. Therefore, we found no evidence that immunotherapy induced tuberculosis occurrence or reactivation. Nevertheless, clinical data of large sample size is still necessary in the future for further verification.

Another important clinical question is when is the best timing to initiate anti-PD-(L)1 therapy in the context of active tuberculosis. ICI therapy was discontinued, temporarily interrupted, continued, or not specified when patients received anti-tuberculosis treatment in previous reported cases [17, 26, 28, 35, 36]. In our study, anti-tuberculosis therapy is mainly started about 4 months before immunotherapy. This treatment paradigm was similar to the previous anti-tuberculosis and anti-tumor therapies, so as to achieve sputum conversion via anti-tuberculosis treatment for 2 to 3 months before anti-tumor treatment. However, to date, there is still no relevant guideline to clarify the order and interval between anti-tuberculosis and anti-tumor treatments.

This study has several limitations. First, the relatively short time for follow-up precludes meaningful survival analysis, and further toxicities may emerge over time. Second, it is a retrospective study. Thus, further prospective clinical studies are required to clarify the indications and management of immunotherapy for patients with coexisting tuberculosis and malignancy.

This study demonstrated that patients with coexisting malignant tumor and tuberculosis showed relatively higher overall response and benefit equally among three groups from anti-PD-(L)1 therapy. Furthermore, anti-tuberculosis treatment was well-controlled for those with active tuberculosis. Notably, the combination of anti-PD-(L)1 and anti-tuberculosis therapy was well-tolerated without unexpected toxic effects. Our results provide the clinical evidence for the application of PD-(L)1 inhibitors in cancer patients with tuberculosis, rendering more available treatment options for these patients. Clinicians may judiciously consider the anti-PD-(L)1 treatment in patients with malignancy and tuberculosis.