This is the first report to focus on a comparison of irSC and irHepatitis. The present study showed that the use of anti-CTLA-4 inhibitor did not induce irSC, and that irSC showed a slow and gradual onset with a ≥grade 2 ALP elevation resulting in a cholestatic pattern, compared with irHepatitis. These findings are clinically important, because a ≥grade 2 ALP elevation resulting in a cholestatic pattern will lead to diagnostic doubt of irSC, suggesting the addition of CT scans.
Distinguishing between irSC and irHepatitis is important for predicting steroid responsiveness. The response rate for steroid therapy is reportedly only 11.5% among irSC patients [11]. Also, in the present study, an irSC case treated with a corticosteroid showed a worsening of bile duct strictures, even after an improvement in the liver enzyme levels [15]. In contrast, 98% of irHepatitis patients showed a resolution of their condition after treatment with corticosteroids or improved without requiring corticosteroids in a previous report [16].
The incidence of irHepatitis depends on the class of ICI and is higher for patients treated with ipilimumab alone (4.5%) or combination with nivolumab (13%), compared with anti-PD-1 inhibitor alone (1.8%) [1, 17]. In contrast, a review of case studies showed that none of the 31 irSC cases were associated with anti-CTLA-4 inhibitor.11 From a histological perspective, anti-CTLA-4 inhibitor causes granulomas hepatitis with central vein endotheliitis, while anti-PD-1/PD-L1 inhibitor induces a heterogenous injury pattern without granulomatous inflammation [18, 19]. Endotheliitis is often present in acute liver allograft rejection [20], in which rejection is induced by the blockage of immunosuppressive regulatory T cells (Tregs) via the CTLA-4 pathway [21]. Anti-CTLA-4 inhibitor binds to CTLA-4 on Tregs in vivo [22], and Treg differentiation is induced by endothelial cells [23]. Anti-CTLA-4 inhibitors might modulate the Treg signaling pathway, inducing irHepatitis with endotheliitis. While PD-1/L-1 inhibitors causes inflammation with a high rate of CD8+/CD4+ cells, which is a pathological feature of immune-related adverse event in the hepatobiliary system [24]. The CD8+ cells associate with sclerosing cholangitis by producing interferon gamma [25]. The invasion of the CD8+ inflammation cells may have influenced its association with the immune-related cholangitis. Different classes of ICIs might have different mechanisms of action and be associated with the development of irSC and irHepatitis.
In the present study, irSC showed a slow and gradual onset with cholestatic pattern compared to irHepatitis. The median number of cycles until the onset of hepatotoxicity has been reported to be 5.5 in irSC patients and 1 to 3 for irHepatitis patients [11, 16]. Thus, the onset of irHepatitis is thought to be more acute than that of irSC.
Overlap is a condition in which multiple diseases coexist, such as autoimmune hepatitis and primary sclerosing cholangitis. The possibility of irHepatitis-irSC overlap should also be considered, as Stuart et al. preciously reported one case [26]. In 16 patients who underwent a liver biopsy out of 536 irHepatitis patients, more than half of them presented with bile duct injury [18]. Conversely, irHepatitis was found in 13.3% of irSC patients [11]. Although a liver biopsy can make a definitive diagnosis of irHepatitis-irSC overlap, bleeding after a liver biopsy is generally considered the major complication. Therefore, bile duct imaging can be an important step in evaluating the coexistence of irSC which indicates a poor responsiveness to corticosteroid and in determining if a liver biopsy is needed even after a diagnosis of irHepatitis.
Clinical factors predicting a risk of immune-related hepatotoxicity include the following: prior autoimmune disease [27], prior immune-related adverse events from ICIs [28], high dose of ICIs, and the combination of ipilimumab and nivolumab [16]. In the present study, a logistic regression analysis showed that the use of anti-CTLA-4 inhibitor was an independent risk factor for ≥grade 3 hepatotoxicity, and that a female sex and melanoma were confounding factors. Because anti-CTLA-4 inhibitor is associated with hepatitis, specific predictors of irSC factors could not be detected in this study.
This study had several limitations. First, the total number of patients with irSC or irHepatitis was small because of a lack of CT, and the unclassified hepatotoxicity remained in some cases. Therefore, a statistical analysis was not done for comparison between irSC and irHepatitis. The reason for the lack of CT data is that a standardized diagnostic strategy for irSC does not yet exist. According to the current guidelines for immunotherapy, all immune-related hepatotoxicity will be regarded as irHepatitis. Therefore, we propose that patients with a ≥grade 2 ALP elevation that cannot be ascribed to other causes should be considered for close follow-up and radiological examination for the diagnosis of irSC. Early detection of irSC will lead to ICI discontinuation and/or initiation of corticosteroid therapy. Second, there is a selection bias in patients who underwent CT scans. In this study, CT scans were performed in 85% (11/13) of patients with a ≥grade 3 AST or ALT elevations, while 3.6% (1/28) of patients with a grade 2 elevation. However, it is grade 3 or higher toxicities that requires immediate initiation of corticosteroids. Third, hepatotoxicity was graded by evaluating the AST/ALT values using the CTCAE system, and not according to the international normalized ratio and the presence of ascites or encephalopathy using the Drug Induced Liver Injury Network severity index, which is suitable for evaluating the severity of hepatotoxicity [29]. The CTCAE system is usually used to evaluate hepatotoxicity in oncology clinical trials and in management guidelines for immune-related adverse events [4, 5, 12]. Future prospective studies are needed to confirm our findings.
In conclusion, immune-related hepatotoxicity includes both irHepatitis and irSC, which have different characteristics such as the class of ICI that has been used and the onset pattern. Distinguishing between irHepatitis and irSC can help to predict the steroid response. Clinicians should be aware that an ALP elevation resulting in a cholestatic pattern might indicate the emergence of irSC, although the current management strategy for immunotherapy is based only on the AST, ALT, and total bilirubin values. This issue should be evaluated and validated in a large-scale clinical study.