RESUMEN
Background: Cardiac surgery-associated acute kidney injury (CSA-AKI) is a major complication that results in short- and long-term mortality among patients. Here, we adopted machine learning algorithms to build prediction models with the overarching goal of identifying patients who are at a high risk of such unfavorable kidney outcomes. Methods: A total of 1686 patients (development cohort) and 422 patients (validation cohort), with 126 pre- and intra-operative variables, were recruited from the First Medical Centre and the Sixth Medical Centre of Chinese PLA General Hospital in Beijing, China, respectively. Analyses were performed using six machine learning techniques, namely K-nearest neighbor, logistic regression, decision tree, random forest (RF), support vector machine, and neural network, and the APPROACH score, a previously established risk score for CSA-AKI. For model tuning, optimal hyperparameter was achieved by using GridSearch with 5-fold cross-validation from the scikit-learn library. Model performance was externally assessed via the receiver operating characteristic (ROC) and decision curve analysis (DCA). Explainable machine learning was performed using the Python SHapley Additive exPlanation (SHAP) package and Seaborn library, which allow the calculation of marginal contributory SHAP value. Results: 637 patients (30.2%) developed CSA-AKI within seven days after surgery. In the external validation, the RF classifier exhibited the best performance among the six machine learning techniques, as shown by the ROC curve and DCA, while the traditional APPROACH risk score showed a relatively poor performance. Further analysis found no specific causative factor contributing to the development of CSA-AKI; rather, the development of CSA-AKI appeared to be a complex process resulting from a complex interplay of multiple risk factors. The SHAP summary plot illustrated the positive or negative contribution of RF-top 20 variables and extrapolated risk of developing CSA-AKI at individual levels. The Seaborn library showed the effect of each single feature on the model output of the RF prediction. Conclusions: Efficient machine learning approaches were successfully established to predict patients with a high probability of developing acute kidney injury after cardiac surgery. These findings are expected to help clinicians to optimize treatment strategies and minimize postoperative complications. Clinical Trial Registration: The study protocol was registered at the ClinicalTrials Registration System (https://www.clinicaltrials.gov/, #NCT04966598) on July 26, 2021.
RESUMEN
BACKGROUND: Immunotherapy has transformed cancer treatment patterns for advanced hepatocellular carcinoma (aHCC) in recent years. Therefore, the identification of predictive biomarkers has important clinical implications. METHODS: We collected medical records from 117 aHCC patients treated with anti-PD-1 antibody. Kaplan-Meier analysis and Cox proportional hazard regression were used to evaluate the association between peripheral blood biomarkers and overall survival (OS) and progression-free survival (PFS). Finally, the prognostic nomogram was constructed. RESULTS: The mPFS and mOS were 7.0 months and 18.7 months, respectively. According to Kaplan-Meier analysis and Cox regression analysis, we regarded the treatment regimen (p = 0.020), hemoglobin (Hb) at 6-week (p = 0.042), neutrophil-to-lymphocyte ratio (NLR) at 6-week (p < 0.001), system immune inflammation index (SII) at 6-week (p = 0.125) as predictors of PFS, and alpha fetoprotein (AFP) (p = 0.035), platelet-to-lymphocyte ratio (PLR) (p = 0.012), Hb at 6-week (p = 0.010) and NLR at 6-week (p = 0.020) as predictors of OS. Furthermore, the results suggest that the OS and PFS nomogram model were in agreement with actual observations. CONCLUSION: Biomarkers in peripheral blood can predict the prognosis of patients with aHCC treated with anti-PD-1 antibody. The development of nomogram models can help us to screen potential patients who can benefit from immunotherapy.
Asunto(s)
Carcinoma Hepatocelular , Neoplasias Hepáticas , Humanos , Carcinoma Hepatocelular/tratamiento farmacológico , Inhibidores de Puntos de Control Inmunológico/uso terapéutico , Nomogramas , Neoplasias Hepáticas/tratamiento farmacológico , Pronóstico , Linfocitos , Biomarcadores , Neutrófilos , Estudios RetrospectivosRESUMEN
The immune effect induced by photodynamic therapy (PDT) has a limited effect on breast tumor. This study hypothesized that suppressive immune checkpoints on T cells might upregulate after PDT, which may reduce the antitumor effect of PDT for treating breast tumor. This study explored the alteration of immune checkpoint for the first time. A bilateral subcutaneous transplanted breast tumor mice model was established, and right tumors imitated primary tumors, and left tumors imitated distant tumors. Primary tumors were treated with PDT mediated by hematoporphyrin derivatives (HpD-PDT). Costimulatory molecules (ICOS, OX40, and 4-1BB) and immune checkpoints (PD1, LAG-3, CTLA-4, TIM-3, TIGIT) on tumor infiltrating T cells after HpD-PDT were analyzed by flow cytometry. Antitumor and immune effects were also assessed after HpD-PDT combined with anti-PD1 and LAG-3 antibodies. Primary tumors were suppressed, but distant tumors could not be inhibited after HpD-PDT. The number of T cells was increased, but function did not enhance after HpD-PDT. Additionally, costimulatory molecules (ICOS, OX40, and 4-1BB) were not elevated, but the suppressive immune checkpoints on tumor infiltrating T cells were upregulated after HpD-PDT. Notably, PD1+ LAG-3+ CD4+ T and PD1+ LAG-3+ CD8+ T cells were significantly increased. When PD1 and LAG-3 blockade combined with HpD-PDT, both primary and distant tumors were significantly suppressed, and antitumor immune effects were significantly enhanced. HpD-PDT could upregulate the PD1+ LAG-3+ CD4+ T and PD1+ LAG-3+ CD8+ T cells. Dual blockade of PD1 and LAG-3 immune checkpoints can enhance the antitumor effect of HpD-PDT.
Asunto(s)
Neoplasias de la Mama , Fotoquimioterapia , Animales , Ratones , Humanos , Femenino , Regulación hacia Arriba , Linfocitos T CD8-positivos , Derivado de la Hematoporfirina , Neoplasias de la Mama/tratamiento farmacológicoRESUMEN
Introduction: Currently, the incidence of liver cancer is on the rise annually. Precise identification of liver tumors is crucial for clinicians to strategize the treatment and combat liver cancer. Thus far, liver tumor contours have been derived through labor-intensive and subjective manual labeling. Computers have gained widespread application in the realm of liver tumor segmentation. Nonetheless, liver tumor segmentation remains a formidable challenge owing to the diverse range of volumes, shapes, and image intensities encountered. Methods: In this article, we introduce an innovative solution called the attention connect network (AC-Net) designed for automated liver tumor segmentation. Building upon the U-shaped network architecture, our approach incorporates 2 critical attention modules: the axial attention module (AAM) and the vision transformer module (VTM), which replace conventional skip-connections to seamlessly integrate spatial features. The AAM facilitates feature fusion by computing axial attention across feature maps, while the VTM operates on the lowest resolution feature maps, employing multihead self-attention, and reshaping the output into a feature map for subsequent concatenation. Furthermore, we employ a specialized loss function tailored to our approach. Our methodology begins with pretraining AC-Net using the LiTS2017 dataset and subsequently fine-tunes it using computed tomography (CT) and magnetic resonance imaging (MRI) data sourced from Hubei Cancer Hospital. Results: The performance metrics for AC-Net on CT data are as follows: dice similarity coefficient (DSC) of 0.90, Jaccard coefficient (JC) of 0.82, recall of 0.92, average symmetric surface distance (ASSD) of 4.59, Hausdorff distance (HD) of 11.96, and precision of 0.89. For AC-Net on MRI data, the metrics are DSC of 0.80, JC of 0.70, recall of 0.82, ASSD of 7.58, HD of 30.26, and precision of 0.84. Conclusion: The comparative experiments highlight that AC-Net exhibits exceptional tumor recognition accuracy when tested on the Hubei Cancer Hospital dataset, demonstrating highly competitive performance for practical clinical applications. Furthermore, the ablation experiments provide conclusive evidence of the efficacy of each module proposed in this article. For those interested, the code for this research article can be accessed at the following GitHub repository: https://github.com/killian-zero/py_tumor-segmentation.git.
Asunto(s)
Neoplasias Hepáticas , Tomografía Computarizada por Rayos X , Humanos , Imagen por Resonancia Magnética , Neoplasias Hepáticas/diagnóstico por imagen , Instituciones Oncológicas , Suministros de Energía Eléctrica , Procesamiento de Imagen Asistido por ComputadorRESUMEN
Purpose: This study aims to investigate the process of small cell lung cancer (SCLC) patients from achieving optimal efficacy to experiencing disease progression until death. It examines the predictive value of the treatment response on progression free survival (PFS) and overall survival (OS) of SCLC patients. Patients and Methods: We conducted a retrospective analysis on 136 SCLC patients diagnosed from 1992 to 2018. Important prognostic factors were identified to construct nomogram models. The predictive performance of the models was evaluated using the receiver operating characteristic curves and calibration curves. Survival differences between groups were compared using Kaplan-Meier survival curves. Subsequently, an independent cohort consisting of 106 SCLC patients diagnosed from 2014 to 2021 was used for validation. Results: We constructed two nomograms to predict first-line PFS (PFS1) and OS of SCLC. The area under the receiver operating characteristic curves for the PFS1 nomogram predicting PFS at 3-, 6-, and 12-months were 0.919 (95% CI: 0.867-0.970), 0.908 (95% CI: 0.860-0.956) and 0.878 (95% CI: 0.798-0.958), and for the OS nomogram predicting OS at 6-, 12-, and 24-months were 0.814 (95% CI: 0.736-0.892), 0.819 (95% CI: 0.749-0.889) and 0.809 (95% CI: 0.678-0.941), indicating those two models with a high discriminative ability. The calibration curves demonstrated the models had a high degree of consistency between predicted and observed values. According to the risk scores, patients were divided into high-risk and low-risk groups, showing a significant difference in survival rate. And these findings were validated in another independent validation cohort. Conclusion: Based on the patients' treatment response after standardized treatment, we developed and validated two nomogram models to predict PFS1 and OS of SCLC. The models demonstrated good accuracy, reliability and clinical applicability by validating in an independent cohort.
RESUMEN
Objective. Radiation therapy (RT) represents a prevalent therapeutic modality for head and neck (H&N) cancer. A crucial phase in RT planning involves the precise delineation of organs-at-risks (OARs), employing computed tomography (CT) scans. Nevertheless, the manual delineation of OARs is a labor-intensive process, necessitating individual scrutiny of each CT image slice, not to mention that a standard CT scan comprises hundreds of such slices. Furthermore, there is a significant domain shift between different institutions' H&N data, which makes traditional semi-supervised learning strategies susceptible to confirmation bias. Therefore, effectively using unlabeled datasets to support annotated datasets for model training has become a critical issue for preventing domain shift and confirmation bias.Approach. In this work, we proposed an innovative cross-domain orthogon-based-perspective consistency (CD-OPC) strategy within a two-branch collaborative training framework, which compels the two sub-networks to acquire valuable features from unrelated perspectives. More specifically, a novel generative pretext task cross-domain prediction (CDP) was designed for learning inherent properties of CT images. Then this prior knowledge was utilized to promote the independent learning of distinct features by the two sub-networks from identical inputs, thereby enhancing the perceptual capabilities of the sub-networks through orthogon-based pseudo-labeling knowledge transfer.Main results. Our CD-OPC model was trained on H&N datasets from nine different institutions, and validated on the four local intuitions' H&N datasets. Among all datasets CD-OPC achieved more advanced performance than other semi-supervised semantic segmentation algorithms.Significance. The CD-OPC method successfully mitigates domain shift and prevents network collapse. In addition, it enhances the network's perceptual abilities, and generates more reliable predictions, thereby further addressing the confirmation bias issue.
Asunto(s)
Aprendizaje Profundo , Neoplasias de Cabeza y Cuello , Humanos , Semántica , Tomografía Computarizada por Rayos X , Neoplasias de Cabeza y Cuello/diagnóstico por imagen , Neoplasias de Cabeza y Cuello/radioterapia , Órganos en Riesgo , Procesamiento de Imagen Asistido por Computador/métodosRESUMEN
Background: Pancreatic ductal adenocarcinoma (PDAC) is in urgent need of a second-line or later-line treatment strategy. We aimed to analyze the efficacy and safety of additional anlotinib, specifically anlotinib in combination with immunotherapy, in patients with PDAC who have failed first-line therapy. Methods: Patients with pathological diagnosis of PDAC were additionally treated with anlotinib, and some patients were treated with anti-PD-1 agents at the same time, which could be retrospectively analyzed. The efficacy and safety of additional anlotinib were evaluated. Results: A total of 23 patients were included. In patients treated with additional anlotinib, the overall median progression-free survival (PFS) was 1.8 months and the median overall survival (OS) was 6.3 months, regardless of anti-PD-1 agents. Among patients receiving additional anlotinib in combination with anti-PD-1 agents, median PFS and OS were 1.8 and 6.5 months, respectively. Adverse events (AEs) were observed in 16 patients (69.6%). In patients treated with additional anlotinib, the majority of AEs were grade 1-3. Univariate analysis revealed that patients with baseline red blood cell distribution width (RDW) <14% treated with additional anlotinib plus anti-PD-1 agents had significantly longer OS than patients with baseline RDW ≥14% (p = 0.025). Patients with additional anlotinib plus anti-PD-1 agents as second-line therapy had a longer OS than those treated as later-line therapy (p = 0.012). Multivariate analysis showed that baseline RDW was the only independent risk factor for OS (p = 0.042). Conclusion: The combination of anlotinib and immunotherapy represents an effective add-on therapy with tolerable AEs as second- or later-line therapy in patients with PDAC, particularly in patients with baseline RDW <14%.
RESUMEN
Background: Recurrent metastasis after radical resection in patients of colorectal cancer (CRC) is a great challenge for the world, in which genomic alterations play a major role in tumorigenesis. MUC4 plays a significant role in recurrence and metastasis in tumor. This study is aimed at exploring the association between MUC4 variants and metastatic recurrence of CRC. Methods: Forty-seven patients relapsing with metastasis and 37 patients remaining disease-free postoperatively were enrolled. Next-generation sequencing (NGS) detected mutations. Mutation and mRNA expression data were downloaded from TCGA and cBioPortal databases. We analyzed the relationship between MUC4 variants and clinical parameters, as well as possible molecular mechanisms. Results: MUC4 variants rs56359992 and rs781124621 were associated with survival in patients with CRC. Rs56359992 was more common in patients with metastatic recurrence. MAPK pathway, PI3K-Akt pathway, JAK-STAT pathway, cell cycle, WNT pathway and mTOR pathway were found to correlate with MUC4 mutation by GO/KEGG analysis, as well as resting and activated mast cell related to MUC4 mutation by CIBERSORT analysis. Conclusion: Genetic variants of MUC4 with CRC may constitute a molecular signature of metastatic recurrence. MUC4 may become a new target for the treatment of CRC recurrence.
RESUMEN
BACKGROUND: Hepatic ischemia-reperfusion (I/R) injury is the main cause of morbidity and mortality after hepatectomy; thus, new methods for reducing I/R injury are required. The aim of this study is to evaluate changes in the average apparent diffusion coefficient (ADCavg) and fractional anisotropy (FA) in rabbits with partial hepatic I/R injury with magnetic resonance diffusion tensor imaging (DTI). METHODS: The left lobe of the rabbit liver underwent 60 minutes of ischemia followed by 0.5, 2, 6, 12, 24, and 48 hours of reperfusion. T2-weighted images (T2WI), T1-weighted images (T1WI), DTI, and contrast-enhanced T1WI were performed; 6 b values were used for DTI on 6 diffusion directions. The serum levels of transaminases and liver histopathology findings were examined. RESULTS: In the early stage of I/R (0.5 hour), ADCavg decreased significantly and increased sharply to 2 hours, then increased from 6 hours to 48 hours of reperfusion, except for a transient decrease (24 hours). Meanwhile, FA showed almost the opposite trend, drastically increasing during the first 0.5 hour and then slightly decreasing until 48 hours of reperfusion, except for an obvious decrease in the 2-hours group. The serum levels of liver markers and the pathologic scores were sharply increased in the I/R group after reperfusion and correlated with DTI of hepatic tissue after I/R. CONCLUSIONS: Diffusion tensor imaging is feasible for imaging I/R-induced liver damage and can discriminate isotropic properties of the liver after I/R injury with objective changes in the ADCavg and FA. Diffusion tensor imaging can be a promising novel approach for use in clinical management after liver surgery.
Asunto(s)
Imagen de Difusión Tensora , Daño por Reperfusión , Animales , Conejos , Imagen de Difusión Tensora/métodos , Imagen por Resonancia Magnética/métodos , Hígado/patología , Daño por Reperfusión/diagnóstico por imagen , Daño por Reperfusión/etiología , Daño por Reperfusión/patología , Isquemia/patología , Espectroscopía de Resonancia MagnéticaRESUMEN
BACKGROUND: Due to the poor prognosis of gastric cancer (GC), early detection methods are urgently needed. Plasma exosomal circular RNAs (circRNAs) have been suggested as novel biomarkers for GC. AIM: To identify a novel biomarker for early detection of GC. METHODS: Healthy donors (HDs) and GC patients diagnosed by pathology were recruited. Nine GC patients and three HDs were selected for exosomal whole-transcriptome RNA sequencing. The expression profiles of circRNAs were analyzed by bioinformatics methods and validated by droplet digital polymerase chain reaction. The expression levels and area under receiver operating characteristic curve values of plasma exosomal circRNAs and standard serum biomarkers were used to compare their diagnostic efficiency. RESULTS: There were 303 participants, including 240 GC patients and 63 HDs, involved in the study. The expression levels of exosomal hsa_circ_0079439 were significantly higher in GC patients than in HDs (P < 0.0001). However, the levels of standard serum biomarkers were similar between the two groups. The area under the curve value of exosomal hsa_circ_0079439 was higher than those of standard biomarkers, including carcinoembryonic antigen, carbohydrate antigen (CA)19-9, CA72-4, alpha-fetoprotein, and CA125 (0.8595 vs 0.5862, 0.5660, 0.5360, 0.5082, and 0.5018, respectively). The expression levels of exosomal hsa_circ_0079439 were significantly decreased after treatment (P < 0.05). Moreover, the expression levels of exosomal hsa_circ_0079439 were obviously higher in early GC (EGC) patients than in HDs (P < 0.0001). CONCLUSION: Our results suggest that plasma exosomal hsa_circ_0079439 is upregulated in GC patients. Moreover, the levels of exosomal hsa_circ_0079439 could distinguish EGC and advanced GC patients from HDs. Therefore, plasma exosomal hsa_circ_0079439 might be a potential biomarker for the diagnosis of GC during both the early and late stages.
Asunto(s)
Neoplasias Gástricas , Humanos , Neoplasias Gástricas/diagnóstico , Neoplasias Gástricas/genética , Detección Precoz del Cáncer , ARN Circular , Antígeno CA-19-9 , Biología ComputacionalRESUMEN
Background: Triple-negative breast cancer (TNBC) is the most aggressive type of breast cancer. Immune checkpoint inhibitors (ICIs) have been widely used to treat various tumors and have changed the landscape of tumor management, but the data from real-world studies of ICIs for TNBC treatment remain limited. The aim of this study was to evaluate the efficacy of ICIs in the treatment of patients with advanced TNBC in a real-world setting and to explore possible correlates. Methods: The clinical data of advanced TNBC patients who received ICI treatment in the Chinese People's Liberation Army (PLA) General Hospital were collected. Treatment responses, outcomes and adverse events (AEs) were assessed. Results: Eighty-one patients were included in the study. The confirmed objective response rate (ORR) was 32.1%, and the disease control rate (DCR) was 64.2%. The median progression-free survival (PFS) was 4.2 months, and the median overall survival (OS) was 11.0 months. PFS and OS were longer in patients who achieved clinical benefit from ICIs and shorter in patients who received later-line ICIs and higher levels of inflammation; specifically, patients with higher TILs had longer PFS. Overall AEs were tolerable. Conclusions: ICIs are effective in the treatment of advanced TNBC, and the adverse reactions are tolerable. A panel of biomarkers including LDH, ALP, and bNLR were identified to predict the efficacies of ICIs in TNBC treatment.
RESUMEN
Background: We developed machine learning models that combine preoperative and intraoperative risk factors to predict mortality after cardiac surgery. Methods: Machine learning involving random forest, neural network, support vector machine, and gradient boosting machine was developed and compared with the risk scores of EuroSCORE I and II, Society of Thoracic Surgeons (STS), as well as a logistic regression model. Clinical data were collected from patients undergoing adult cardiac surgery at the First Medical Centre of Chinese PLA General Hospital between December 2008 and December 2017. The primary outcome was post-operative mortality. Model performance was estimated using several metrics, including sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC). The visualization algorithm was implemented using Shapley's additive explanations. Results: A total of 5,443 patients were enrolled during the study period. The mean EuroSCORE II score was 3.7%, and the actual in-hospital mortality rate was 2.7%. For predicting operative mortality after cardiac surgery, the AUC scores were 0.87, 0.79, 0.81, and 0.82 for random forest, neural network, support vector machine, and gradient boosting machine, compared with 0.70, 0.73, 0.71, and 0.74 for EuroSCORE I and II, STS, and logistic regression model. Shapley's additive explanations analysis of random forest yielded the top-20 predictors and individual-level explanations for each prediction. Conclusions: Machine learning models based on available clinical data may be superior to clinical scoring tools in predicting postoperative mortality in patients following cardiac surgery. Explanatory models show the potential to provide personalized risk profiles for individuals by accounting for the contribution of influencing factors. Additional prospective multicenter studies are warranted to confirm the clinical benefit of these machine learning-driven models.
RESUMEN
Neutrophil extracellular traps (NETs), products of neutrophil death when exposed to certain stimuli, were first proposed as a type of response to bacterial infection in infectious diseases. Since then, extensive studies have discovered its involvement in other non-infectious inflammatory diseases including thromboembolism, autoimmune diseases, and cancer. Colorectal cancer (CRC) is one of the most common malignancies in the world. NET formation is closely associated with tumorigenesis, progression, and metastasis in CRC. Therefore, the application of NETs in clinical practice as diagnostic biomarkers, therapeutic targets, and prognostic predictors has a promising prospect. In addition, therapeutics targeting NETs are significantly efficient in halting tumor progression in preclinical cancer models, which further indicates its potential clinical utility in cancer treatment. This review focuses on the stimuli of NETosis, its pro-tumorigenic activity, and prospective clinical utility primarily in but not limited to CRC.
RESUMEN
Autophagy is a self-degradative pathway by which subcellular elements are broken down intracellularly to maintain cellular homeostasis. Cardiac autophagy commonly decreases with aging and is accompanied by the accumulation of misfolded proteins and dysfunctional organelles, which are undesirable to the cell. Reduction of autophagy over time leads to aging-related cardiac dysfunction and is inversely related to longevity. However, despite the increasing interest in autophagy in cardiac diseases and aging, the process remains an undervalued and disregarded object in calcific valvular disease. Neither the nature through which autophagy is triggered nor the interplay between autophagic machinery and targeted molecules during aortic valve calcification are fully understood. Recently, the upregulation of autophagy has been shown to result in cardioprotective effects against cell death as well as its origin. Here, we review the evidence that shows how autophagy can be both beneficial and detrimental as it pertains to aortic valve calcification in the heart.
RESUMEN
BACKGROUND: Combination therapy with immune checkpoint inhibitors (ICIs) and antiangiogenic agents is generally effective and well tolerated and might be effective for metastatic urothelial carcinoma (UC). However, ICI treatment is often associated with unique responses, such as pseudoprogression and ICI-related pneumonitis (CIP), which may influence clinical decision making and affect treatment. Although there have been many studies on the mechanism of pseudoprogression and CIP, the characteristics and relationship of these special events in a clinical setting remain rarely reported. CASE PRESENTATION: Here, we present a patient with lung metastatic UC who underwent surgery and two lines of chemotherapy. The programmed cell death-1 (PD-1) inhibitor nivolumab and antiangiogenics agent bevacizumab were used as maintenance treatments. The patient experienced pseudoprogression after 2 PD-1 inhibitor cycles. The lesions in both lungs were enlarged on computed tomography (CT) imaging, and treatments were continued for another two cycles, after which the tumor size decreased to below baseline, followed by a durable response. However, after 4 months of pseudoprogression, the patient then developed CIP. The CIP was responsive to glucocorticoid therapy but recurred during ICI rechallenge, leading to the termination of immune therapy. Ultimately, the patient achieved durable, stable disease for over 18 months without further anticancer treatment. CONCLUSIONS: Our case shows that pseudoprogression can occur in UC during immunotherapy even when combined with an effective antiangiogenic agent. In addition, pseudoprogression may be correlated with future adverse effects and a durable response. In the management of CIP, early rechallenge with ICIs may lead to CIP recurrence, which could be more severe and needs to be treated early and with appropriate drugs. Clinicians should be aware of atypical responses to ICIs and adjust the treatment plan accordingly.