Automated image analysis of NSCLC biopsies to predict response to anti-PD-L1 therapy

Low Infiltration of CD8+ PD-L1+ T Cells and M2 Macrophages Predicts Improved Clinical Outcomes After Immune Checkpoint Inhibitor Therapy in Non-Small Cell Lung Carcinoma

Background

Many clinical studies have shown that patients with non-small cell lung carcinoma (NSCLC) can benefit from immune checkpoint inhibitor (ICI) therapy; however, PD-L1 and tumor mutation burden (TMB), which are recommended by the NCCN guidelines, are still insufficient in predicting the response to and prognosis of immunotherapy. Given the widespread use of ICIs, it is important to find biomarkers that can predict immunotherapy outcomes in NSCLC patients, and the exploration of additional effective biomarkers for ICI therapy is urgently needed.

Methods

A total of 33 stage II-IV NSCLC patients were included in this study. We analyzed immune markers in biopsy and surgical tissue resected from these patients before treatment with ICIs. We examined the infiltration of immune cells and expression of PD-L1 in immune cells using fluorescent multiplex immunohistochemistry (mIHC) stained with CD8/CD68/CD163/PD-L1 antibodies.

Results

In this cohort, we observed that the levels of CD8+ T cells, CD8+PD-L1+ T cells, and CD68+CD163+ M2 macrophages in the total region were independent prognostic factors for progression-free survival (PFS) in NSCLC patients treated with ICIs (HR=0.04, P=0.013; HR=17.70, P=0.026; and HR=17.88, P=0.011, respectively). High infiltration of CD8+ T cells and low infiltration of CD8+PD-L1+ T cells throughout the region were correlated with prolonged PFS (P=0.016 and P=0.02, respectively). No statistically significant difference was observed for CD68+CD163+ M2 macrophages. The joint parameters CD8+ high/CD8+PD-L1+ low, CD8+ high/CD68+CD163+ low and CD8+PD-L1+ low/CD68+CD163+ low predicted better PFS than other joint parameters (P<0.01, P<0.01, and P<0.001, respectively), and they also demonstrated stronger stratification than single biomarkers. The response rate of patients with high infiltration of CD8+ T cells was significantly higher than that of those with low infiltration (P<0.01), and the joint parameters CD8+/CD8+PD-L1+ and CD8+/CD68+CD163+ also demonstrated stronger stratification than single biomarkers.

Conclusions

This retrospective study identified the predictive value of CD8+PD-L1+ T cells, CD8+ T cells, and CD68+CD163+ M2 macrophages in NSCLC patients who received ICIs. Interestingly, our results indicate that the evaluation of joint parameters has certain significance in guiding ICI treatment in NSCLC patients.

Multiplex Immunofluorescence and Multispectral Imaging: Forming the Basis of a Clinical Test Platform for Immuno-Oncology

As immuno-oncology (I/O) emerges as an effective approach in the fight against cancer, multispectral imaging of multiplex immunofluorescence (mIF) is maturing as an analytical platform. The timing is fortuitous. Due to health economic considerations surrounding the use of I/O, there is an urgent need for tests that accurately predict response to the growing list of available therapies. Multispectral mIF provides several advantages over other biomarker modalities by enabling deeper interrogation of the intricate biology within the tumor microenvironment, including detection of cell-to-cell spatial interactions that correlate with clinical outcomes. It also provides a practical path for generating reliable and reproducible results in a clinically suitable, high-throughput workflow. In this article, we (1) describe the principles behind multispectral mIF; (2) provide advice and recommendations on assay development and optimization and highlight characteristics of a well-performing assay; and (3) discuss the requirements for translating this approach into clinical practice.

Dual-scale categorization based deep learning to evaluate programmed cell death ligand 1 expression in non-small cell lung cancer

ABSTRACT

In precision oncology, immune check point blockade therapy has quickly emerged as novel strategy by its efficacy, where programmed death ligand 1 (PD-L1) expression is used as a clinically validated predictive biomarker of response for the therapy. Automating pathological image analysis and accelerating pathology evaluation is becoming an unmet need. Artificial Intelligence and deep learning tools in digital pathology have been studied in order to evaluate PD-L1 expression in PD-L1 immunohistochemistry image. We proposed a Dual-scale Categorization (DSC)-based deep learning method that employed 2 VGG16 neural networks, 1 network for 1 scale, to critically evaluate PD-L1 expression. The DSC-based deep learning method was tested in a cohort of 110 patients diagnosed as non-small cell lung cancer. This method showed a concordance of 88% with pathologist, which was higher than concordance of 83% of 1-scale categorization-based method. Our results show that the DSCbased method can empower the deep learning application in digital pathology and facilitate computer-aided diagnosis.

Intratumoral CD39+CD8+ T Cells Predict Response to Programmed Cell Death Protein-1 or Programmed Death Ligand-1 Blockade in Patients With NSCLC

INTRODUCTION

Programmed cell death protein-1 (PD-1) and programmed death-ligand 1 (PD-L1) blockade is currently widely used in the treatment of metastatic NSCLC. Despite available biomarker stratification, clinical responses vary. Thus, the search for novel biomarkers with improved response prediction is ongoing. Previously, using mass cytometry or cytometry by time-of-flight (CyTOF), our group demonstrated that CD39⁺CD8⁺ immune cells represent tumor antigen-specific, cytotoxic T cells in treatment-naive NSCLC. We hypothesized that accurate quantitation of this T cell subset would predict immunotherapy outcome.

METHODS

To translate this to a clinical setting, the present study compared CyTOF data with a range of clinically relevant methods, including conventional immunohistochemistry (IHC), multiplex IHC or immunofluorescence (mIHC), and gene expression assay by NanoString.

RESULTS

Quantification using mIHC but not conventional IHC or NanoString correlated with the CyTOF results. The specificity and sensitivity of mIHC were then evaluated in a separate retrospective NSCLC cohort. CD39⁺CD8⁺ T cell proportion, as determined by mIHC, successfully stratified responders and nonresponders to PD-1 or PD-L1 inhibitors (objective response rate of 63.6%, compared with 0% for the negative group). This predictive capability was independent from other confounding factors, such as total CD8⁺ T cell proportion, CD39⁺ lymphocyte proportion, PD-L1 positivity, EGFR mutation status, and other clinicopathologic parameters.

CONCLUSIONS

Our results suggest that the mIHC platform is a clinically relevant method to evaluate CD39⁺CD8⁺ T cell proportion and that this marker can serve as a potential biomarker that predicts response to PD-1 or PD-L1 blockade in patients with NSCLC. Further validation in additional NSCLC cohorts is warranted.

Malignant Pleural Effusions-A Window Into Local Anti-Tumor T Cell Immunity?

The success of immunotherapy that targets inhibitory T cell receptors for the treatment of multiple cancers has seen the anti-tumor immune response re-emerge as a promising biomarker of response to therapy. Longitudinal characterization of T cells in the tumor microenvironment (TME) helps us understand how to promote effective anti-tumor immunity. However, serial analyses at the tumor site are rarely feasible in clinical practice. Malignant pleural effusions (MPE) associated with thoracic cancers are an abnormal accumulation of fluid in the pleural space that is routinely drained for patient symptom control. This fluid contains tumor cells and immune cells, including lymphocytes, macrophages and dendritic cells, providing a window into the local tumor microenvironment. Recurrent MPE is common, and provides an opportunity for longitudinal analysis of the tumor site in a clinical setting. Here, we review the phenotype of MPE-derived T cells, comparing them to tumor and blood T cells. We discuss the benefits and limitations of their use as potential dynamic biomarkers of response to therapy.

Next-Generation Digital Histopathology of the Tumor Microenvironment

Progress in cancer research is substantially dependent on innovative technologies that permit a concerted analysis of the tumor microenvironment and the cellular phenotypes resulting from somatic mutations and post-translational modifications. In view of a large number of genes, multiplied by differential splicing as well as post-translational protein modifications, the ability to identify and quantify the actual phenotypes of individual cell populations in situ, i.e., in their tissue environment, has become a prerequisite for understanding tumorigenesis and cancer progression. The need for quantitative analyses has led to a renaissance of optical instruments and imaging techniques. With the emergence of precision medicine, automated analysis of a constantly increasing number of cellular markers and their measurement in spatial context have become increasingly necessary to understand the molecular mechanisms that lead to different pathways of disease progression in individual patients. In this review, we summarize the joint effort that academia and industry have undertaken to establish methods and protocols for molecular profiling and immunophenotyping of cancer tissues for next-generation digital histopathology-which is characterized by the use of whole-slide imaging (brightfield, widefield fluorescence, confocal, multispectral, and/or multiplexing technologies) combined with state-of-the-art image cytometry and advanced methods for machine and deep learning.

Artificial Intelligence & Tissue Biomarkers: Advantages, Risks and Perspectives for Pathology

Tissue Biomarkers are information written in the tissue and used in Pathology to recognize specific subsets of patients with diagnostic, prognostic or predictive purposes, thus representing the key elements of Personalized Medicine. The advent of Artificial Intelligence (AI) promises to further reinforce the role of Pathology in the scenario of Personalized Medicine: AI-based devices are expected to standardize the evaluation of tissue biomarkers and also to discover novel information, which would otherwise be ignored by human review, and use them to make specific predictions. In this review we will present how AI has been used to support Tissue Biomarkers evaluation in the specific field of Pathology, give an insight to the intriguing field of AI-based biomarkers and discuss possible advantages, risk and perspectives for Pathology.

Anti-PD-(L)1 immunotherapy for brain metastases in non-small cell lung cancer: Mechanisms, advances, and challenges

The brain is one of the most common metastatic sites in non-small cell lung cancer (NSCLC), which is associated with an extremely poor prognosis. Despite the availability of several therapeutic options, the treatment efficacy remains unsatisfactory for NSCLC brain metastases. Anti-programmed cell death-1 (PD-1) and its ligand (PD-L1) monoclonal antibodies have reshaped therapeutic strategies in advanced NSCLC. Preliminary evidence has shown that anti-PD-(L)1 monotherapy is also effective in NSCLC patients with brain metastases. However, the traditional view asserted that these therapeutic antibodies were incapable of crossing the blood-brain barrier (BBB) with large molecular size, thus most patients with brain metastases were excluded from most studies on anti-PD-(L)1 immunotherapy. Therefore, the efficacy and its mechanisms of action of anti-PD-(L)1 immunotherapy against brain metastases in NSCLC have not been clarified. In this review, we will survey the underlying mechanisms and current clinical advances of anti-PD-(L)1 immunotherapy in the treatment of brain metastases in NSCLC. The trafficking of activated cytotoxic T cells that are mainly derived from the primary tumor and deep cervical lymph nodes is critical for the intracranial response to anti-PD-(L)1 immunotherapy, which is driven by interferon-γ (IFN-γ). Additionally, promising combined strategies with the rationale in the treatment of brain metastases will be presented to provide future directions for clinical study design. Several significant challenges in the preclinical and clinical studies of brain metastases, as well as potential solutions, will also be discussed.

SITC cancer immunotherapy resource document: a compass in the land of biomarker discovery

Since the publication of the Society for Immunotherapy of Cancer's (SITC) original cancer immunotherapy biomarkers resource document, there have been remarkable breakthroughs in cancer immunotherapy, in particular the development and approval of immune checkpoint inhibitors, engineered cellular therapies, and tumor vaccines to unleash antitumor immune activity. The most notable feature of these breakthroughs is the achievement of durable clinical responses in some patients, enabling long-term survival. These durable responses have been noted in tumor types that were not previously considered immunotherapy-sensitive, suggesting that all patients with cancer may have the potential to benefit from immunotherapy. However, a persistent challenge in the field is the fact that only a minority of patients respond to immunotherapy, especially those therapies that rely on endogenous immune activation such as checkpoint inhibitors and vaccination due to the complex and heterogeneous immune escape mechanisms which can develop in each patient. Therefore, the development of robust biomarkers for each immunotherapy strategy, enabling rational patient selection and the design of precise combination therapies, is key for the continued success and improvement of immunotherapy. In this document, we summarize and update established biomarkers, guidelines, and regulatory considerations for clinical immune biomarker development, discuss well-known and novel technologies for biomarker discovery and validation, and provide tools and resources that can be used by the biomarker research community to facilitate the continued development of immuno-oncology and aid in the goal of durable responses in all patients.

The immune contexture and Immunoscore in cancer prognosis and therapeutic efficacy

The international American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC) tumour-node-metastasis (TNM) staging system provides the current guidelines for the classification of cancer. However, among patients within the same stage, the clinical outcome can be very different. More recently, a novel definition of cancer has emerged, implicating at all stages a complex and dynamic interaction between tumour cells and the immune system. This has enabled the definition of the immune contexture, representing the pre-existing immune parameters associated with patient survival. Even so, the role of distinct immune cell types in modulating cancer progression is increasingly emerging. An immune-based assay named the 'Immunoscore' was defined to quantify the in situ T cell infiltrate and was demonstrated to be superior to the AJCC/UICC TNM classification for patients with colorectal cancer. This Review provides a broad overview of the main immune parameters positively or negatively shaping cancer development, including the Immunoscore, and their prognostic and predictive value. The importance of the immune system in cancer control is demonstrated by the requirement for a pre-existing intratumour adaptive immune response for effective immunotherapies, such as checkpoint inhibitors. Finally, we discuss how the combination of multiple immune parameters, rather than individual ones, might increase prognostic and/or predictive power.

A narrative review of digital pathology and artificial intelligence: focusing on lung cancer

The emergence of whole slide imaging technology allows for pathology diagnosis on a computer screen. The applications of digital pathology are expanding, from supporting remote institutes suffering from a shortage of pathologists to routine use in daily diagnosis including that of lung cancer. Through practice and research large archival databases of digital pathology images have been developed that will facilitate the development of artificial intelligence (AI) methods for image analysis. Currently, several AI applications have been reported in the field of lung cancer; these include the segmentation of carcinoma foci, detection of lymph node metastasis, counting of tumor cells, and prediction of gene mutations. Although the integration of AI algorithms into clinical practice remains a significant challenge, we have implemented tumor cell count for genetic analysis, a helpful application for routine use. Our experience suggests that pathologists often overestimate the contents of tumor cells, and the use of AI-based analysis increases the accuracy and makes the tasks less tedious. However, there are several difficulties encountered in the practical use of AI in clinical diagnosis. These include the lack of sufficient annotated data for the development and validation of AI systems, the explainability of black box AI models, such as those based on deep learning that offer the most promising performance, and the difficulty in defining the ground truth data for training and validation owing to inherent ambiguity in most applications. All of these together present significant challenges in the development and clinical translation of AI methods in the practice of pathology. Additional research on these problems will help in resolving the barriers to the clinical use of AI. Helping pathologists in developing knowledge of the working and limitations of AI will benefit the use of AI in both diagnostics and research.

CLPTM1L induces estrogen receptor β signaling-mediated radioresistance in non-small cell lung cancer cells

INTRODUCTION

Radioresistance is a major challenge in lung cancer radiotherapy, and new radiosensitizers are urgently needed. Estrogen receptor β (ERβ) is involved in the progression of non-small cell lung cancer (NSCLC), however, the role of ERβ in the response to radiotherapy in lung cancer remains elusive. In the present study, we investigated the mechanism underlying ERβ-mediated transcriptional activation and radioresistance of NSCLC cells.

METHODS

Quantitative real-time PCR, western blot and immunohistochemistry were used to detect the expression of CLPTM1L, ERβ and other target genes. The mechanism of CLPTM1L in modulation of radiosensitivity was investigated by chromatin immunoprecipitation assay, luciferase reporter gene assay, immunofluorescence staining, confocal microscopy, coimmunoprecipitation and GST pull-down assays. The functional role of CLPTM1L was detected by function assays in vitro and in vivo.

RESULTS

CLPTM1L expression was negatively correlated with the radiosensitivity of NSCLC cell lines, and irradiation upregulated CLPTM1L in radioresistant (A549) but not in radiosensitive (H460) NSCLC cells. Meanwhile, IR induced the translocation of CLPTM1L from the cytoplasm into the nucleus in NSCLC cells. Moreover, CLPTM1L induced radioresistance in NSCLC cells. iTRAQ-based analysis and cDNA microarray identified irradiation-related genes commonly targeted by CLPTM1L and ERβ, and CLPTM1L upregulated ERβ-induced genes CDC25A, c-Jun, and BCL2. Mechanistically, CLPTM1L coactivated ERβ by directly interacting with ERβ through the LXXLL NR (nuclear receptor)-binding motif. Functionally, ERβ silencing was sufficient to block CLPTM1L-enhanced radioresistance of NSCLC cells in vitro. CLPTM1L shRNA treatment in combination with irradiation significantly inhibited cancer cell growth in NSCLC xenograft tumors in vivo.

CONCLUSIONS

The present results indicate that CLPTM1L acts as a critical coactivator of ERβ to promote the transcription of its target genes and induce radioresistance of NSCLC cells, suggesting a new target for radiosensitization in NSCLC therapy. Video Abstract.

Predictive biomarkers for cancer immunotherapy with immune checkpoint inhibitors

Although the clinical development of immune checkpoint inhibitors (ICIs) therapy has ushered in a new era of anti-tumor therapy, with sustained responses and significant survival advantages observed in multiple tumors, most patients do not benefit. Therefore, more and more attention has been paid to the identification and development of predictive biomarkers for the response of ICIs, and more in-depth and comprehensive understanding has been continuously explored in recent years. Predictive markers of ICIs efficacy have been gradually explored from the expression of intermolecular interactions within tumor cells to the expression of various molecules and cells in tumor microenvironment, and been extended to the exploration of circulating and host systemic markers. With the development of high-throughput sequencing and microarray technology, a variety of biomarker strategies have been deeply explored and gradually achieved the process from the identification of single marker to the development of multifactorial synergistic predictive markers. Comprehensive predictive-models developed by integrating different types of data based on different components of tumor-host interactions is the direction of future research and will have a profound impact in the field of precision immuno-oncology. In this review, we deeply analyze the exploration course and research progress of predictive biomarkers as an adjunctive tool to tumor immunotherapy in effectively identifying the efficacy of ICIs, and discuss their future directions in achieving precision immuno-oncology.

Development and applications of computer image analysis algorithms for scoring of PD-L1 immunohistochemistry

Immune checkpoint inhibitors targeting programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) have rapidly become integral to standard-of-care therapy for non-small cell lung cancer and other cancers. Immunohistochemical (IHC) staining of PD-L1 is currently the accepted and approved diagnostic assay for selecting patients for PD-L1/PD-1 axis therapies in certain indications. However, the inherent biological complexity of PD-L1 and the availability of several PD-L1 assays – each with different detection systems, platforms, scoring algorithms and cut-offs – have created challenges to ensure reliable and reproducible results based on subjective visual assessment by pathologists. The increasing adoption of computer technologies into the daily workflow of pathology provides an opportunity to leverage these tools towards improving the clinical value of PD-L1 IHC assays. This review describes several image analysis software programs of computer-aided PD-L1 scoring in the hope of driving further discussion and technological advancement in digital pathology and artificial intelligence approaches, particularly as precision medicine evolves to encompass accurate simultaneous assessment of multiple features of cancer cells and their interactions with the tumor microenvironment.

Comparison of continuous measures across diagnostic PD-L1 assays in non-small cell lung cancer using automated image analysis

Tumor programmed cell death ligand-1 (PD-L1) expression is a key biomarker to identify patients with non-small cell lung cancer who may have an enhanced response to anti-programmed cell death-1 (PD-1)/PD-L1 treatment. Such treatments are used in conjunction with PD-L1 diagnostic immunohistochemistry assays. We developed a computer-aided automated image analysis with customized PD-L1 scoring algorithm that was evaluated via correlation with manual pathologist scores and used to determine comparability across PD-L1 immunohistochemistry assays. The image analysis scoring algorithm was developed to quantify the percentage of PD-L1 positive tumor cells on scans of whole-slide images of archival tumor samples from commercially available non-small cell lung cancer cases, stained with four immunohistochemistry PD-L1 assays (Ventana SP263 and SP142 and Dako 22C3 and 28-8). The scans were co-registered and tumor and exclusion annotations aligned to ensure that analysis of each case was restricted to comparable tissue areas. Reference pathologist scores were available from previous studies. F1, a statistical measure of precision and recall, and overall percentage agreement scores were used to assess concordance between pathologist and image analysis scores and between immunohistochemistry assays. In total, 471 PD-L1-evalulable samples were amenable to image analysis scoring. Image analysis and pathologist scores were highly concordant, with F1 scores ranging from 0.8 to 0.9 across varying matched PD-L1 cutoffs. Based on F1 and overall percentage agreement scores (both manual and image analysis scoring), the Ventana SP263 and Dako 28-8 and 22C3 assays were concordant across a broad range of cutoffs; however, the Ventana SP142 assay showed very different characteristics. In summary, a novel automated image analysis scoring algorithm was developed that was highly correlated with pathologist scores. The algorithm permitted quantitative comparison of existing PD-L1 diagnostic assays, confirming previous findings that indicate a high concordance between the Ventana SP263 and Dako 22C3 and 28-8 PD-L1 immunohistochemistry assays.

CTLA-4 Immunohistochemistry and Quantitative Image Analysis for Profiling of Human Cancers

There is an important need in immuno-oncology to develop reliable immunohistochemistry (IHC) to assess the expression of CTLA-4⁺ tumor-infiltrating lymphocytes in human cancers and quantify them with image analysis (IA). We used commercial polyclonal and monoclonal antibodies and characterized three chromogenic cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) assays with suitable specificity and sensitivity for use in formalin-fixed, paraffin-embedded (FFPE) tissues. We found variable numbers of CTLA-4⁺ lymphocytes in multiple types of cancer and secondary lymphoid organs (SLOs) and other normal human tissues. Combining CTLA-4 with CD3, CD4, or CD8 by immunofluorescence showed that CTLA-4⁺ lymphocytes in SLOs and tumors were typically CD3⁺ and CD4⁺, but not CD8⁺. Individual lymphocytes expressed CTLA-4 either as primarily granular cytoplasmic staining or as excentric globular deposits. The CTLA-4/FoxP3 (forkhead box P3 protein) duplex IHC demonstrated that CTLA-4⁺/FoxP3^- lymphocytes predominated in the germinal centers of SLOs and tumor tertiary lymphoid structures (TLSs), whereas CTLA-4⁺/FoxP3⁺ lymphocytes populated the T-cell zone of SLOs and TLSs, plus tumor stroma. IA scoring was highly comparable with pathologist scoring for CTLA-4 and CTLA-4/FoxP3 assays and a FoxP3 single IHC. Our findings show that CTLA-4 IHC can be used to reliably label lymphocytes in FFPE human tissues, making it possible to investigate the role of CTLA-4 in the tumor microenvironment.

Comparison of Biomarker Modalities for Predicting Response to PD-1/PD-L1 Checkpoint Blockade: A Systematic Review and Meta-analysis

IMPORTANCE

PD-L1 (programmed cell death ligand 1) immunohistochemistry (IHC), tumor mutational burden (TMB), gene expression profiling (GEP), and multiplex immunohistochemistry/immunofluorescence (mIHC/IF) assays have been used to assess pretreatment tumor tissue to predict response to anti-PD-1/PD-L1 therapies. However, the relative diagnostic performance of these modalities has yet to be established.

OBJECTIVE

To compare studies that assessed the diagnostic accuracy of PD-L1 IHC, TMB, GEP, and mIHC/IF in predicting response to anti-PD-1/PD-L1 therapy.

EVIDENCE REVIEW

A search of PubMed (from inception to June 2018) and 2013 to 2018 annual meeting abstracts from the American Association for Cancer Research, American Society of Clinical Oncology, European Society for Medical Oncology, and Society for Immunotherapy of Cancer was conducted to identify studies that examined the use of PD-L1 IHC, TMB, GEP, and mIHC/IF assays to determine objective response to anti-PD-1/PD-L1 therapy. For PD-L1 IHC, only clinical trials that resulted in US Food and Drug Administration approval of indications for anti-PD-1/PD-L1 were included. Studies combining more than 1 modality were also included. Preferred Reporting Items for Systematic Reviews and Meta-analysis guidelines were followed. Two reviewers independently extracted the clinical outcomes and test results for each individual study.

MAIN OUTCOMES AND MEASURES

Summary receiver operating characteristic (sROC) curves; their associated area under the curve (AUC); and pooled sensitivity, specificity, positive and negative predictive values (PPV, NPV), and positive and negative likelihood ratios (LR+ and LR-) for each assay modality.

RESULTS

Tumor specimens representing over 10 different solid tumor types in 8135 patients were assayed, and the results were correlated with anti-PD-1/PD-L1 response. When each modality was evaluated with sROC curves, mIHC/IF had a significantly higher AUC (0.79) compared with PD-L1 IHC (AUC, 0.65, P < .001), GEP (AUC, 0.65, P = .003), and TMB (AUC, 0.69, P = .049). When multiple different modalities were combined such as PD-L1 IHC and/or GEP + TMB, the AUC drew nearer to that of mIHC/IF (0.74). All modalities demonstrated comparable NPV and LR-, whereas mIHC/IF demonstrated higher PPV (0.63) and LR+ (2.86) than the other approaches.

CONCLUSIONS AND RELEVANCE

In this meta-analysis, tumor mutational burden, PD-L1 IHC, and GEP demonstrated comparable AUCs in predicting response to anti-PD-1/PD-L1 treatment. Multiplex immunohistochemistry/IF and multimodality biomarker strategies appear to be associated with improved performance over PD-L1 IHC, TMB, or GEP alone. Further studies with mIHC/IF and composite approaches with a larger number of patients will be required to confirm these findings. Additional study is also required to determine the most predictive analyte combinations and to determine whether biomarker modality performance varies by tumor type.