Search

Objectives: This study aims to identify prognostic factors associated with metastatic recurrence-free survival of cervical carcinoma (CC) patients treated with radical radiotherapy and assess the cure probability of radical radiotherapy from metastatic recurrence. Methods: Data were from 446 cervical carcinoma patients with radical radiotherapy for an average follow up of 3.96 years. We applied a mixture cure model to investigate the association between metastatic recurrence and prognostic factors and the association between noncure probability and factors, respectively. A nonparametric test of cure probability under the framework of a mixture cure model was used to examine the significance of cure probability of the definitive radiotherapy treatment. Propensity-score-matched (PSM) pairs were generated to reduce bias in subgroup analysis. Results: Patients in advanced stages (p = 0.005) and those with worse treatment responses in the 3rd month (p = 0.004) had higher metastatic recurrence rates. Nonparametric tests of the cure probability showed that 3-year cure probability from metastatic recurrence was significantly larger than 0, and 5-year cure probability was significantly larger than 0.7 but no larger than 0.8. The empirical cure probability by mixture cure model was 79.2% (95% CI: 78.6–79.9%) for the entire study population, and the overall median metastatic recurrence time for uncured patients (patients susceptible to metastatic recurrence) was 1.60 (95% CI: 1.51–1.69) years. Locally advanced/advanced stage was a risk factor but non-significant against the cure probability (OR = 1.078, p = 0.088). The interaction of age and activity of radioactive source were statistically significant in the incidence model (OR = 0.839, p = 0.025). In subgroup analysis, compared with high activity of radioactive source (HARS), low activity of radioactive source (LARS) significantly contributed to a 16.1% higher cure probability for patients greater than 53 years old, while cure probability was 12.2% lower for the younger patients. Conclusions: There was statistically significant evidence in the data showing the existence of a large amount of patients cured by the definitive radiotherapy treatment. HARS is a protective factor against metastatic recurrence for uncured patients, and young patients tend to benefit more than the elderly from the HARS treatment.

Inter-well Fracture Driven Interaction (FDI), aka Frac Hit, has gained visibility as more high intensity events (pressure increase in parent wells). Severe production performance impact (parent and child wells) are observed in depleted and tightly spaced wells in unconventional reservoirs. Various strategies, including water pre-loading in parent wells, adjusting well completion design of child wells, optimizing development plans, etc., have been tried in the industry to mitigate the negative impact of FDIs on production performance. However, mixed results have been observed in some pilot tests utilizing a water pre-loading mitigation strategy. Hence, it is critical to understand the driving mechanisms and key parameters impacting pre-loading performance in order to quantify the expected benefit of pre-loading in different well scenarios and optimize the pre-loading design.A novel integrated and coupled geomechanics / fluid flow / fracturing propagation modeling workflow is applied to simulate the FDI (including pre-loading) on the pad scale. The workflow is capable of simulating the entire life cycle of the FDI interaction, (a) stimulation of parent wells, (b) production of parent well, (c) pre-loading of the parent well, (d) stimulation of the child wells, and (e) production forecasting of both parent and child wells. The workflow includes calibration with field observations of fracture geometry, production performance, pre-loading pressure behavior, etc. Geomechanics and reservoir simulation coupling captures the stress change due to the depletion at the parent well and re-pressurizing from pre-loading.The workflow was applied to one pre-loading pilot test in the Permian Basin. The goal of the study is to identify the key parameters that contribute to the pre-loading success and optimize the pre-loading design. History match of parent and child wells production performance using the integrated and coupled geomechanics / fluid flow / fracturing propagation modeling workflow was conducted to calibrate and validate the model. Then, a sensitivity study was performed on various key parameters, including pre-loading volume and level of depletion at the parent well to understand its impact on the production performance at the parent and child wells. Learnings from this study provided guidance to the operation team in order to optimize FDI mitigation strategy.

Background: Detection of genomic structural variants (SVs) in breast cancer was of great clinical interest. Numerous methods were developed to call SVs from third-generation sequencing data, but the results and accuracy of various calling methods differ greatly. Thus, assessing the accuracy of these calling methods became important and challenging especially in the absence of gold standards. This study aimed to compare the performance of two SVs calling methods, denoted by T1 and T2 respectively, without gold standards of SVs. Methods: A total of 155,566 structural variants detected by the two calling methods with third-generation sequencing were collected based on 20 breast cancer patients from Peking University People's Hospital. Utilizing a Bayesian approach that could take account of the dependency between calling results, we assessed the accuracy of T1 and T2 in detecting insertion, deletion, duplication, inversion, translocation, and copy number variant, respectively, in terms of sensitivity and specificity. Sensitivity analysis was conducted to evaluate the robustness of the comparison results. Results: The estimated sensitivity and specificity of T2 were higher than that of T1 in detecting different SVs. The estimated sensitivity of T2 in detecting insertion, deletion, and duplication were 0.9438, 0.9440, and 0.9441, respectively, while the sensitivity of T1 were 0.8750, 0.8743, and 0.8843, respectively. Sensitivity analysis showed that the prior information is critical but the comparison results were robust in a certain range. Conclusions: Results showed that T2 significantly outperforms T1 in a general comparison. The proposed model and obtained results are helpful to evaluate and select better SVs calling methods in the absence of gold standards, which in turn will be conducive to identifying new SVs associated with breast cancer or other diseases.

Androgen receptor (AR) expression is frequently observed in breast cancer, but its association with estrogen receptor (ER) expression in breast cancer remains unclear. This study analyzed the clinicopathological and molecular features associated with AR negativity in both ER-positive and ER-negative breast cancer, trying to elucidate the molecular correlation between AR and ER. Our results showed that AR negativity was associated with different clinicopathological characteristics and molecular features in ER-positive and ER-negative breast cancer. Moreover, AR-positive breast cancer has better clinicopathological features than AR-negative breast cancer, especially in the ER-negative subtype. These results suggest that the role of AR in ER-negative breast cancer is distinctive from that in ER-positive breast cancer.

This research was carried out to quantify the duration from symptom onset to recovery/death (SOR/SOD) during the first four waves and the Alpha/Delta period of the epidemic in Khyber Pakhtunkhwa, Pakistan, and identify the associated factors. A total of 173,894 COVID-19 cases were admitted between 16 March 2020 and 30 November 2021, including 458 intensive care unit (ICU) cases. The results showed that the case fatality rate (CFR) increased with age, and females had a higher CFR. The median SOR of ICU cases was longer than that of non-ICU cases (27.6 vs. 17.0 days), while the median SOD was much shorter (6.9 vs. 8.4 days). The SOR and SOD in the Delta period were slightly shortened than the Alpha period. Age, cardiovascular diseases, chronic lung disease, diabetes, fever, breathing issues, and ICU admission were risk factors that were significantly associated with SOD (p < 0.001). A control measure, in-home quarantine, was found to be significantly associated with longer SOD (odds ratio = 9.49, p < 0.001). Infected vaccinated individuals had longer SOD than unvaccinated individuals, especially for cases that had received two vaccine doses (p < 0.001). Finally, an advice on getting full-dose vaccination is given specifically to individuals aged 20–59 years.

Currently, a large number of mature methods are available for gaze estimation. However, most regular gaze estimation approaches require additional hardware or platforms with professional equipment for data collection or computing that typically involve high costs and are relatively tedious. Besides, the implementation is particularly complex. Traditional gaze estimation approaches usually require systematic prior knowledge or expertise for practical operations. Moreover, they are primarily based on the characteristics of pupil and iris, which uses pupil shapes or infrared light and iris glint to estimate gaze, requiring high-quality images shot in special environments and other light source or professional equipment. We herein propose a two-stage gaze estimation method that relies on deep learning methods and logistic regression, which can be applied to various mobile platforms without additional hardware devices or systematic prior knowledge. A set of automatic and fast data collection mechanism is designed for collecting gaze images through a mobile platform camera. Additionally, we propose a new annotation method that improves the prediction accuracy and outperforms the traditional gridding annotation method. Our method achieves good results and can be adapted to different applications.

BackgroundHER2-low-positive breast cancers are reported to have distinct clinical and molecular features from HER2-zero tumors in ER-positive breast cancer. However, the association between HER2-status with response to endocrine therapy is largely unknown. MethodsIn this study, we conducted a retrospective analysis to compare the response of ER-positive breast cancer to neoadjuvant endocrine therapy (NET), neoadjuvant chemotherapy (NCT) and survival according to HER2-status.ResultsPathological complete response (pCR) in primary tumor (ypT0/isNx) is significantly lower in HER2-low-positive breast cancer than in HER2-zero tumors for NET cohort (18 out of 233 [7.7%] vs 9 out of 45 [20%]) but not for NCT cohort (28 out of 204 [13.3%] vs 5 out of 36 [13.9%]). HER2-zero tumors consistently show higher ypT0/isNx pCR ratio than HER2-low-positive breast cancer in all endocrine therapy duration groups including 3-10months, 11-20months and 21-30months groups. Multi-variable logistic regression analyses showed that HER2-low-positive expression is a risk factor for poor response to NET. In patients receiving NET, HER2-low-positve breast tumors has better disease-free survival compared with HER2-zero while no significance difference was detected with overall survival. ConclusionsOur results demonstrated that HER2-low-positive breast tumors was a distinct subtype in ER-positive breast cancers and was associated with endocrine therapy resistance.

Regression analysis of heterogeneous samples with subgroup structure is essential to the development of precision medicine. In practice, this task is often challenging owing to the lack of prior knowledge of subgroup labels. Therefore, detecting the subgroups with similar characteristics becomes critical, which often controls the accuracy of regression analysis. In this article, we investigate a new framework for detecting the subgroups that have similar characters in feature space and similar treatment effects. The key idea is that we incorporate K-means clustering into the regression framework of concave pairwise fusion, so that the regression and subgroup detection tasks can be performed simultaneously. Our method is specifically tailored for handling the situations where the sample is not homogeneous in the sense that the response variables in different domains of feature space are generated through different mechanisms.

Background: Structural variations (SVs) are common genetic alterations in the human genome that could cause different phenotypes and various diseases including cancer. However, the detection of structural variations using the second-generation sequencing was limited by its short read-length which in turn restrained our understanding of structural variations. Methods: In this study, we developed a 28-gene panel for long-read sequencing and employed it to both Oxford Nanopore Technologies and Pacific Biosciences platforms. We analyzed structural variations in the 28 breast cancer-related genes through long-read genomic and transcriptomic sequencing of tumor, para-tumor and blood samples in 19 breast cancer patients. Results: Our results showed that some somatic SVs were recurring among the selected genes, though the majority of them occurred in the non-exonic region. We found evidence supporting the existence of hotspot regions for SVs, which extended our previous understanding that they exist only for single nucleotide variations. Conclusions: In conclusion, we employed long-read genomic and transcriptomic sequencing in identifying SVs from breast cancer patients and proved that this approach holds great potential in clinical application.

Background To evaluate the diagnostic value of combinations of tumor markers carcinoembryonic antigen (CEA), carbohydrate antigen (CA) 125, CA153, and CA19-9 in identifying malignant pleural effusion (MPE) from non-malignant pleural effusion (non-MPE) using machine learning, and compare the performance of popular machine learning methods. Methods A total of 319 samples were collected from patients with pleural effusion in Beijing and Wuhan, China, from January 2018 to June 2020. Five machine learning methods including Logistic regression, extreme gradient boosting (XGBoost), Bayesian additive regression tree, random forest, and support vector machine were applied to evaluate the diagnostic performance. Sensitivity, specificity, Youden's index, and the area under the receiver operating characteristic curve (AUC) were used to evaluate the performance of different diagnostic models. Results For diagnostic models with a single tumor marker, the model using CEA, constructed by XGBoost, performed best (AUC = 0.895, sensitivity = 0.80), and the model with CA153, also by XGBoost, showed the largest specificity 0.98. Among all combinations of tumor markers, the combination of CEA and CA153 achieved the best performance (AUC = 0.921, sensitivity = 0.85) in identifying MPE under the diagnostic model constructed by XGBoost. Conclusions Diagnostic models for MPE with a combination of multiple tumor markers outperformed the models with a single tumor marker, particularly in sensitivity. Using machine learning methods, especially XGBoost, could comprehensively improve the diagnostic accuracy of MPE.

Objective: to evaluate the risk of treatment-related adverse events of different severity and different system with PD-1 or PD-L1 inhibitors. Methods: randomized controlled trials (RCTs) that using PD-1/PD-L1 for cancer treatment were searched in the PubMed, Embase, Cochrane Library, and Web of Science from 1 January 2019 to 31 May 2021. Adverse events data were extracted from clinical trials website or original article by two authors separately. Meta-analysis was used to determine risk ratio (RR) and 95% confidence interval (95% CI) of adverse events in PD-1/PD-L1 inhibitors groups compared to that of control groups. Subgroup analyses were also performed. Results: a total of 5,807 studies were initially identified and after exclusion, 41 studies were included in meta-analysis. All the trials were international multicenter, randomized, phase II/III clinical trials, with the median follow-up of 27.5 months on average. Analysis of all grade adverse events showed that PD-1/PD-L1 inhibitors treatment significantly increased the risk of immune-related adverse events, including pruritus (RR: 2.34, 95% CI: 1.85–2.96), rash (RR: 1.53, 95% CI: 1.25–1.87), ALT elevation (RR 1.54, 95% CI 1.23–1.92), AST elevation (AST: RR 1.49, 95% CI 1.20–1.85), hepatitis (RR: 3.54, 95% CI: 1.96–6.38) and hypothyroid (RR: 5.29, 95% CI: 4.00–6.99) compared with that of control group. Besides that, PD-1/PD-L1 inhibitors were associated with higher risk of adverse events related to respiratory system including cough (RR: 1.33, 95% CI: 1.21–1.48), dyspnea (RR:1.23, 95% CI: 1.12–1.35) and chest pain (RR: 1.26, 95% CI: 1.07–1.47) compared with that of control groups in our meta-analysis and the dyspnea was taken high risk both in all grade and grade 3 or higher (RR: 1.55, 95% CI: 1.13–2.12). The risk of arthralgia was increased with PD-1/PD-L1 inhibitors (RR: 1.27, 95% CI: 1.10–1.47). Although the risk of myalgia was similar with PD-1/PD-L1 inhibitors and control groups, under subgroup analysis, PD-1/PD-L1 inhibitors decreased the risk of myalgia (RR: 0.56, 95% CI: 0.45–0.70) compared with that of chemotherapy. Conclusions: our results provide clear evidence that the risk of treatment-related adverse events in PD-1 or PD-L1 varies widely in different system. In particular, when using PD-1/PD-L1 inhibitors for oncology treatment, besides the common immune-related adverse events like pruritus, rash, hepatitis, and hypothyroid, the respiratory disorders and musculoskeletal disorders, such as cough, dyspnea, arthralgia, and myalgia, should also be taken into consideration.

Optimum subsurface well spacing is key to developing unconventional reservoirs. From a field application perspective, this paper presents our systematic study on subsurface well spacing in the Permian Basin for unconventional reservoirs which consisted of four main components, i.e., numerical modeling, well interference quantification through simulation and regression, field pilot analogs, and economic evaluation. In this paper, field pilot wells are actual wells in the field to test different well spacing and completion designs. To capture upsized and downsized completions, a wide range of fracture designs in combination with different well spacing scenarios are conducted for both hydraulic fracture modeling and reservoir simulation. At the section level (1 mile by 1 mile, 640 acres), multiple wells are simulated to better capture well interactions and reservoir property variations. A complex fracture network is generated by considering interactions between natural fracture and hydraulic fracture. Well interference, which is determined by estimating ultimate recovery (EUR) difference between a single well and a middle well from multiple wells, is analyzed for general trend regressions and validated through field test results. At the section level, economics is done to evaluate capital efficiency of various scenarios. Results from both modeling work and pilots indicate that larger hydraulic fracture size without an increase in subsurface well spacing does not necessarily improve section EUR and there is a point of diminishing returns. Larger subsurface well spacing with bigger fracture size is thereby a preferable combination.

A comprehensive workflow was developed to support short and long-term unconventional Midland and Delaware Basin development strategy. The workflow is applied to every new pad to ensure child wells are targeting more of the virgin rock. The developed workflow considers pressure and stress changes around parent wells, landing strategy, completion optimization, frac order design, etc. A 3-D reservoir model was developed to estimate the depletion and the induced stress changes around the parent wells. Hydraulic fracture modeling is coupled with the flow simulation model to assess child wells fracture propagation under different scenarios. Different landing strategies were investigated to reduce depletion effects on Child wells. Child wells fracture and proppant fluid intensity was optimized to provide the optimum fracture interference. Certain technologies were successfully utilized to change the pressure and stress around the existing wells to properly alter child well fracture propagation towards virgin rock. Frac order was adjusted accordingly to benefit from the induced changes in reservoir pressure and stress around parent wells. The workflow was applied to areas in the Wolfcamp formation within the Midland and Delaware Basins. Results show the effectiveness of the developed workflow to maintain Basin development performance.

Depletion effects occurs in unconventional reservoirs when the hydraulic fractured wells are completed within the drainage volume of existing producers. Field production and monitoring data show that the existing (parent) wells negatively affect the new (child) wells' productivity, making the new wells produce less than if all the wells are drilled and produced at the same time. This paper presents a systematic study on quantifying the offset well depletion effect in the Permian basin through advanced reservoir modeling and data analytics from field pilots. The workflow starts with building an Earth model for parent wells in the area of interest, generating a hydraulic fracture network, and performing reservoir simulation for production forecast. The depletion effect, including changes in geomechanical properties on child wells, is properly updated in the Earth model and further captured through fracture geometry distortion and production decrease in the depleted environment. The results from the simulation models are further validated with actual field data. Simulations and pilot studies in both Midland and Delaware basins demonstrate that offset well depletion effect can have significant impact on production performance. The ability to quantify depletion effect has significant business impacts on development sequence, facility design, completion design, and overall project economics. Analysis on various development scenarios was performed to evaluate the economic impact of depletion effect and determine how to mitigate this effect. Net present value (NPV) economic analysis of different simulation results indicates that time duration of production in the parent well ahead of child well has more impact on depletion effect than the distances between parent and child wells. As examples, depletion quantifications are applied to optimize well production near lease boundary with parent well depletion on the competitor land and guide section development strategy by evaluating pad sequence scenarios.

Objective High-dose-rate (HDR) afterloading brachytherapy using Iridium-192 source involves large radiation activity varieties due to fast decay. It was unknown but clinically desirable to evaluate its impacts on patient outcomes to support more informed decisions. Methods Data of 510 cervical carcinoma (CC) patients were retrospectively included. High-radioactive (HR) and low-radioactive (LR) groups were statistically defined per patient-specific average mean-dose-rate (MDR) of all fractions. The cutoffs were calculated using R-3.6.1 packages based on significance of correlation with binary outcome or survival time. Categorized 1-month and 3-month follow-up results were analyzed as short-term outcomes. Long-term outcomes were evaluated using local recurrence-free survival (LRFS) and metastatic recurrence-free survival (MRFS). Propensity-score-matched (PSM) pairs were generated to reduce bias. Results The median follow-up time was 47.1 months (interquartile range: 33.9 months–66.4 months), involving MDR varieties of up to 9 folds ranging from 6059.99 cGy/h to 54013.66 cGy/h due to 17 source replacements at intervals ranging from 93 days–199 days. Both short-term (1-month: p = 0.22; 3-month: p = 0.79) and long-term (LRFS: p = 0.10; MRFS: p = 0.46) outcomes showed no significant difference between HR and LR. Subgroup analysis displayed significantly better results in LR for stage I–II (3-month, p = 0.02) and stage II (LRFS, p = 0.04) patients. Both LRFS and MRFS of LR were significantly non-inferior to HR (p ≤ 0.02). Conclusions LR is clinically non-inferior or partially superior to HR for CC treatment using HDR, which dispels concerns of potentially undermined patient outcomes when source replacement is delayed.

Large cohort studies are commonly launched to study the risk effect of genetic variants or other risk factors on a chronic disorder. In these studies, family data are often collected to provide add...

Summary It is important to determine several key parameters, such as well spacing, completions design, landing strategy, and pad sequence, for a successful full-field development of the unconventional reservoir that involves multiple wells and pads in a given area of interest. Those parameters are normally considered individually through small and simple models. In this paper, focusing on developing the whole area effectively, we provided a systematic work flow to handle such challenges together: We first recommended a top-down concept that better represents actual field development and illustrates the importance of the 3D Earth model for the unconventional reservoir; we then proposed an integrated modeling that is an iterative loop consisting of the 3D Earth model, hydraulic-fracture modeling, reservoir simulation, and uncertainty analysis. It is uncommon to build a 3D Earth model for the unconventional reservoir mainly because of the lack of data and software capability. In this paper, we provided a cost-effective approach for the first time on the basis of a large amount of existing vertical wells, newly drilled horizontal wells, and all the data available. A 3D Earth model using information from approximately 1,100 vertical wells from the Midland Basin was presented. Such a model has a high resolution conditioned by high well density, and has an advantage of capturing heterogeneities and interactions more than a simplified model created either from one well or low-resolution seismic interpretation. The model was fed into hydraulic-fracture modeling with the consideration of natural-fracture network and stress shadow, followed by reservoir simulation. The in-house uncertainty-analysis package that functions by experimental-design philosophy is linked to the Earth model, hydraulic-fracture modeling, and reservoir simulation. For the first time, the impacts of all the parameters together were evaluated through the final production performance. In our example, we considered completions design, discrete-fracture-network (DFN) characterization and generation, unpropped hydraulic-fracture properties, fracture compaction, and matrix permeability. The result indicated that DFN characterization is the most important parameter affecting production performance. We applied our model and work flow to field development. Well spacing and pad sequence were studied in this paper as two examples. We demonstrated that it is important to properly consider complex interactions among multiple clusters, stages, and wells to evaluate the impacts on well spacing, completions, and development sequence.

Complex heterogeneous reservoir exhibits special characteristics, such as multiple storage types, wide range of storage sizes, complicated storage combinations, and various flow behaviors. Multi-continuum model is needed to rigorously simulate this type of reservoir. However most reservoir simulators use dual porous media model, and the theory of the models is mainly focused on matrix-fracture system. Therefore when porous media is more complex than matrix-fracture system, such software cannot be utilized to simulate these reservoirs directly. Some studies have been proposed to solve the problem, but most of them are either at the theoretical level or cannot been used widely in the simulation softwares. The goal of this paper is not to develop new numerical simulation software, but to obtain appropriate ways to improve performance of current simulation software for complex reservoir. Our approach is theoretically based on multi-continuum seepage theory. By comparing the microscopic seepage model and multi-continuum seepage model, we find that the conventional theory becomes improper because of the absence of an important restriction to the interporosity flow. In the conventional way, an N-multiple flow model always can be established based on N types of continuous media. In fact, one more restrictive condition needs to be added to make it valid: all the crossflow between any two types of continuum can be simplified into a source or sink terms in the seepage model. Consequently, the multiplicity of the seepage model is always less or equal to the number of the continuum types. In addition, because different type of dual-medium has different crossflow, it is required to construct their crossflow models respectively and derive corresponding calculation formula. Several steps developed in this paper are as follows on the basis of multi-continuum theory and the restrictions appended: First, reservoir storage is categorized into continuous and non-continuous media according to storage scale characteristics, distribution and intensity. In general, small cavities, small fractures, and pores in the rock are continuous while large-scale fractures, big cavities or caves are non-continuous. The whole reservoir is therefore classified into continuum region and dis-continuum region. Second, in the continuum region, multi-media model and its multiplicity are determined by using the new theory proposed in this paper. Third, calculation formulae are constructed for each type of crossflow. This research takes fractured vuggy reservoirs as an example. A new calculation for the flux transfer is derived for such fracture-vug dual-porosity system. Forth, in the discrete region, flow in a filled non-continuous medium is described by Darcy's law and free flow in large unfilled cavities or caves follows Navier-Stokes equation and an approximate approach to using Darcy flow simulation is provided. Lastly, this paper introduces the methods how to simplify multi-media into two media and how to simulate the complex reservoir by using the existing simulators. The proposed method is applied to a naturally fractured carbonate reservoir with cavities in the Tarim Basin in China. The simulation results illustrate the production mechanisms of reservoir development using water flooding, including water injection and shut-in, stairs-shaped line of water cut with time, high production with low water cut for long time.

Programmed cell death receptor-1 and its ligand-1 (PD-1/PD-L1) inhibitors have been applied to many cancers, but the difference of treatment-related adverse events (AEs) across cancer types remains unknown. We performed a meta-analysis and systemic review to compare the incidences of commonly reported all-grade AEs across cancer types and found that the most frequent AEs were fatigue, rash/pruritus, loss of appetite/nausea and diarrhea. However, each cancer type also had its higher incidences of AEs involving a relevant system, such as melanoma with epidermal AEs (rash, diarrhea and enterocolitis), lung cancer with dyspnea and pneumonitis, digestive system cancers with amylase and lipase elevation; and renal cell and urothelial cancer with kidney injury (creatinine elevation and proteinuria). However, the incidence of hepatitis did not follow the pattern to show a difference. We did another comparison between PD-1 and PD-L1 inhibitors in lung cancer and urothelial cancer respectively, and found that the risk of most AEs did not differ much, except for more hypothyroidism in PD-1 inhibitors, and more kidney injury in PD-L1 inhibitors. Besides possible immunological mechanisms for treatment-related AEs, the influence of previous radiotherapy and the clinical characteristics of the diseases themselves should also be considered and is worth further investigation. With the result of this meta-analysis, clinicians could estimate the risk of certain AE in certain cancer type, to make treatment options and to customize monitor strategies.