Risk stratification for nasopharyngeal carcinoma: a real-world study based on locoregional extension patterns and Epstein-Barr virus DNA load

Construction and application of nasopharyngeal carcinoma-specific big data platform based on electronic health records

OBJECTIVE

To establish a nasopharyngeal carcinoma-specific big data platform based on electronic health records (EHRs) to provide data support for real-world study of nasopharyngeal carcinoma.

METHODS

A multidisciplinary expert team was established for this project. Based on industry standards and practical feasibility, the team designed the nasopharyngeal carcinoma data element standards including 14 modules and 640 fields. Data from patients diagnosed with nasopharyngeal carcinoma who visited Southern Hospital after 1999 were extracted from 15 EHRs systems and were cleaned, structured, and standardized using information technologies such as machine learning and natural language processing. In addition, a series of measures such as quality control and data encryption were taken to ensure data quality and patient privacy. At the platform application level, 10 functional modules were designed according to the needs of nasopharyngeal carcinoma research.

RESULTS

As of 1 October 2022, the Big Data platform has included 11,617patients, of whom 8228 (70.83 %) were male and 3389 (29.17 %) were female, with a median age of 48 years (interquartile range, 40 years). The data in the platform were validated to have a high level of completeness and accuracy, especially for key variables such as social demographics, laboratory tests and vital signs. Currently, six projects involving risk factors, early diagnosis, treatment efficacy and prevention of treatment-related toxic reactions have been conducted on the platform.

CONCLUSIONS

We have established a high-quality NPC-specific big data platform by integrating heterogeneous data from multiple sources in the EHR. The platform provides an effective tool and strong data support for real-world studies of nasopharyngeal carcinoma, which helps to improve research efficiency, reduce costs, and improve the quality of research results. We expect to promote multicenter nasopharyngeal carcinoma data sharing in the future to facilitate the generation of high-quality real-world evidence in nasopharyngeal carcinoma. This article may provide some reference value for other comprehensive hospitals to establish a big data platform for nasopharyngeal carcinoma.

Risk stratification of LA-NPC during chemoradiotherapy based on clinical classification and TVRR

PURPOSE

To investigate the correlation between tumor volume reduction rate (TVRR) and prognosis in patients with diverse clinical types of nasopharyngeal carcinoma (NPC) undergoing chemoradiotherapy, thereby aptly categorizing risks and directing the personalized treatment of NPC.

MATERIALS AND METHODS

A total of 605 NPC patients with varying clinical types were enrolled in this study and subsequently segregated into six subgroups based on their clinical types and TVRR. To accentuate the efficacy of grouping, Groups 1-6 underwent clustered analysis of hazard atio (HR) values pertaining to progression-free survival (PFS), forming three risk clusters denoted as low, intermediate, and high. The log-rank test was employed to discern differences, and R 4.1.1 was utilized for cluster analysis.

RESULTS

According to survival rates, we classified the first (G2 and G4), second (G1 and G6), and third (G3 and G5) risk clusters as low-, intermediate-, and high-risk, respectively. When comparing risk stratification with the 8th edition of the TNM staging system, our classification exhibited superior predictive prognostic performance. Subgroup analysis of treatments for each risk cluster revealed that the PFS in the neoadjuvant chemotherapy (NACT) + concurrent chemoradiotherapy (CCRT) group surpassed that of the CCRT group significantly (p < 0.05).

CONCLUSION

The reliance on clinical types and TVRR facilitates risk stratification of NPC during chemoradiotherapy, providing a foundation for physicians to tailor therapeutic strategies. Moreover, the risk cluster delineated for NPC patients during the mid-term of chemoradiotherapy stands as an independent prognostic factor for progression-free survival (PFS), overall survival (OS), distantmetastasis-free survival (DMFS), and local recurrence-free (LRRFS) posttreatment. Additionally, individuals in the high-risk cluster are recommended to undergo adjuvant chemotherapy after CCRT.

The value of integrating tumor volume and plasma Epstein-Barr virus DNA load during sequential chemoradiotherapy for prognostic prediction and therapeutic guidance in high-risk locoregionally advanced nasopharyngeal carcinoma

OBJECTIVES

To investigate the value of integrating primary gross tumor volume (GTVp) and gross tumor volume of nodes (GTVn) after induction chemotherapy (IC) and dynamic changes in plasma cell-free Epstein-Barr virus DNA (cfEBV DNA) during sequential chemoradiotherapy (CRT) in high-risk locoregionally advanced nasopharyngeal carcinoma (LA-NPC).

MATERIALS AND METHODS

We retrospectively reviewed 988 patients with LA-NPC undergoing IC plus concurrent chemoradiotherapy (CCRT) between 2014 and 2018. The entire cohort was divided into four subgroups according to tumor volume and the cfEBV DNA load. Using a supervised statistical clustering approach, we stratified the subgroups into three clusters. Overall survival (OS), disease-free survival (DFS), distant metastasis-free survival (DMFS) and locoregional relapse-free survival (LRRFS) were calculated using Kaplan-Meier analysis and inter-group differences were compared using the log-rank test.

RESULTS

We observed that GTVp & GTVn and cfEBV DNApostIC & cfEBV DNApostCRT were powerful prognostic factors for OS (p = 0.004, p < 0.001, p < 0.001, and p < 0.001, respectively). The survival curves of the three clusters were significantly different. The 5-year OS for the low-risk, intermediate-risk and high-risk clusters were 97.0%, 86.2% and 77.1% (all P values < 0.001), respectively. The risk stratification system showed better predictive performance than the current tumor-node-metastasis (TNM) classification for OS (area under curve [AUC]: 0.653 versus 0.560, p < 0.001), DFS (AUC: 0.639 versus 0.540, p < 0.001), DMFS (AUC: 0.628 versus 0.535, p < 0.001) and LRRFS (AUC: 0.616 versus 0.513, p < 0.001).

CONCLUSION

Both tumor volume and the cfEBV DNA level during sequential CRT are effective prognostic indicators for patients with high-risk LA-NPC. The developed risk stratification system incorporating above factors improved survival prediction and demonstrated potential value in decision-making.

Nomogram Based on Hemoglobin, Albumin, Lymphocyte and Platelet Score to Predict Overall Survival in Patients with T3-4N0-1 Nasopharyngeal Carcinoma

Purpose

There is still uncertainty regarding the prognosis of nasopharyngeal carcinoma (NPC) based on hemoglobin, albumin, lymphocytes, and platelets (HALP) score. The aim of this study was to build and verify a nomogram using HALP score to investigate the prognostic value of NPC and identify low-risk patients in T3-4N0-1 NPC to guide treatment options.

Patients and methods

A total of 568 NPC patients with stage T3-4N0-1M0 were recruited in the study, who were given either concurrent chemoradiotherapy (CCRT) or induction chemotherapy (IC) plus CCRT. The prognostic factors of overall survival (OS) were picked by Cox proportional hazards regression analysis to generate a nomogram, which appraised by discrimination, calibration and clinical utility. Patients were stratified according to risk scores calculated by the nomogram, and compared to the 8th TNM staging system using the Kaplan-Meier methods.

Results

Multivariate analysis showed that TNM stage, Epstein-Barr virus DNA (EBV DNA), HALP score, lactate dehydrogenase-to-albumin ratio (LAR) and systemic inflammatory response index (SIRI) were independent prognostic indicators for OS, and these factors contained in the nomogram. The nomogram demonstrated a significant enhancement over the 8th TNM staging system in terms of assessing OS (C-index, 0.744 vs 0.615 in the training cohort, P < 0.001; 0.757 vs 0.646 in the validation cohort, P = 0.002). Calibration curves displayed good agreement and the stratification in high-risk and low-risk groups resulted in a significant divergence of Kaplan-Meier curves for OS (P < 0.001). In addition, the decision analysis (DCA) curves confirmed satisfactory discriminability and clinical utility.

Conclusion

The HALP score was an independent prognostic factor for NPC. The prognostic function of the nomogram for T3-4N0-1 NPC patients was more accurate compared to the 8th TNM system, facilitating personalized treatment planning.

Cisplatin-based concurrent chemoradiotherapy improved the survival of locoregionally advanced nasopharyngeal carcinoma after induction chemotherapy by reducing early treatment failure

PURPOSE

The aims of this study focusing on Locoregionally advanced nasopharyngeal carcinoma (LANPC) were mainly two-fold: on the one hand, to establish a cut-off value to differentiate early and late failure based on prognosis after recurrence or metastasis; and on the other hand, to investigate the duration of concurrent cisplatin benefit over follow-up time. The results of our study have the potential to guide clinical practice and follow-up.

METHODS

In total, 3123 patients with stage III-IVa NPC receiving Induction chemotherapy followed by concurrent cisplatin or not were analysed. The cut-off value of treatment failure was calculated using the minimum P-value approach. Random survival forest (RSF) model was to simulate the cumulative probabilities of treatment failure (locoregional recurrence and /or distant metastasis) over-time, as well as the monthly time-specific, event-occurring probabilities, for patients at different treatment groups.

RESULTS

Based on subsequent prognosis, early locoregional failure (ELRF) should be defined as recurrence within 14 months (P = 1.47 × 10 - 3), and early distant failure (EDF) should be defined as recurrence within 20 months (P = 1.95 × 10 - 4). A cumulative cisplatin dose (CCD) > 200 mg/m² independently reduced the risk of EDF (hazard ratio, 0.351; 95% confidence interval (CI), 0.169-0.732; P = 0.005). Better failure-free survival (FFS) and overall survival (OS) were observed in concurrent chemotherapy settings ([0 mg/m² vs. 1-200 mg/m² vs. >200 mg/m²]: FFS: 70.4% vs. 74.4% vs. 82.6%, all P < 0.03; OS: 79.5% vs. 83.8% vs. 90.8%, all P < 0.01). In the monthly analysis, treatment failure mainly occurred during the first 4 years, and the risk of distant failure in patients treated with concurrent chemotherapy never exceeded that of patients without concurrent chemotherapy.

CONCLUSION

Locoregional failure that developed within 14 months and/or distant failure within 20 months had poorer subsequent survival. Concurrent chemotherapy provides a significant FFS benefit, primarily by reducing EDF, translating into a long-term OS benefit.

Determining the suitability of definitive radiation therapy in patients with metastatic nasopharyngeal carcinoma based on PET/CT: a large cohort study

OBJECTIVES

To determine patients with de novo metastatic nasopharyngeal carcinoma (mNPC) who would benefit from receiving definitive radiation therapy (DRT) along with their pre-existing palliative chemotherapy (PCT) by evaluating their post-PCT Deauville scores and EBV DNA.

METHODS

A total of 570 mNPC patients, treated with PCT or PCT+DRT, were studied. EBV DNA levels, along with post-PCT Deauville scores, were used to stratify risk based on the recursive partitioning analysis (RPA).

RESULTS

Significant differences were observed in the survival rates of patients with Deauville scores of 1-3 and 4-5 (2-year progression-free survival (PFS): 23.4% versus 8.5%, p < 0.001; 2-year overall survival (OS): 56.8% versus 18.8%, p < 0.001). RPA yielded three distinct groups in the increasing order of risk (Deauville scores of all RPA I-II were within the range of 1-3): (1) RPA I: EBV DNA levels at a pretreatment concentration ≤ 4000 copies/mL and undetectable post-PCT; (2) RPA II: EBV DNA levels either at a pretreatment concentration > 4000 copies/mL or at a pretreatment concentration ≤ 4000 copies/mL and detectable post-PCT; (3) RPA III: Deauville scores 4-5. While patients in RPA I and RPA II had significantly PFS rates when treated with PCT+DRT than when treated with PCT alone (RPA I: 72.7% versus 13.4%, RPA II: 37.8% versus 6.3%), those in RPA III did not experience such PFS benefits (6.5% versus 9.7%).

CONCLUSION

PCT+DRT might improve the survival rates in mNPC patients in the low- and mid-risk strata but not those of patients in the high-risk strata.

KEY POINTS

We use the Deauville scores and the concentrations of the Epstein-Barr virus (EBV) DNA to determine those patients with de novo metastatic NPC who would benefit from radiation therapy.

Prognostic significance of AKR1C4 and the advantage of combining EBV DNA to stratify patients at high risk of locoregional recurrence of nasopharyngeal carcinoma

Background

Distinguishing patients at a greater risk of recurrence is essential for treating locoregional advanced nasopharyngeal carcinoma (NPC). This study aimed to explore the potential of aldo–keto reductase 1C4 (AKR1C4) in stratifying patients at high risk of locoregional relapse.

Methods

A total of 179 patients with locoregionally advanced NPC were grouped by different strategies; they were: (a) divided into two groups according to AKR1C4 expression level, and (b) classified into three clusters by integrating AKR1C4 and Epstein-Barr virus (EBV) DNA. The Kaplan–Meier method was used to calculate locoregional relapse-free survival (LRFS), overall survival (OS), progression-free survival (PFS), and distant metastasis-free survival (DMFS). The Cox proportional hazards model was used to determine potential prognostic factors, and a nomogram was generated to predict 3-year and 5-year LRFS.

Results

A significant difference in the 5-year LRFS was observed between the high and low AKR1C4 expression groups (83.3% vs. 92.7%, respectively; p = 0.009). After integrating AKR1C4 expression and EBV DNA, the LRFS (84.7%, 84.5%, 96.9%, p = 0.014) of high-, intermediate-, and low- AKR1C4 and EBV DNA was also significant. Multivariate analysis indicated that AKR1C4 expression (p = 0.006) was an independent prognostic factor for LRFS. The prognostic factors incorporated into the nomogram were AKR1C4 expression, T stage, and EBV DNA, and the concordance index of the nomogram for locoregional relapse was 0.718.

Conclusions

In conclusion, high AKR1C4 expression was associated with a high possibility of relapse in NPC patients, and integrating EBV DNA and AKR1C4 can stratify high-risk patients with locoregional recurrence.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-022-09924-3.

The prognostic value of adding systemic inflammation response index to Epstein-Barr virus DNA in childhood nasopharyngeal carcinoma: A real-world study

BACKGROUND

To assess the prognostic value of the systemic inflammation response index (SIRI) combined with plasma load of Epstein-Barr virus (EBV) DNA in children and adolescents with locoregionally advanced nasopharyngeal carcinoma (CALANPC).

METHODS

A total of 205 consecutive patients with CALANPC were enrolled. We used recursive partitioning analysis (RPA) to classify patients into various risk groups, with a primary endpoint of overall survival (OS).

RESULTS

Elevated SIRI (≥1.53) and EBV DNA (≥4000 copy/ml) were significantly associated with inferior OS in CALANPC. RPA categorized patients into low- and high-risk groups based on prognostic factors. Survival curves showed excellent discrimination in OS (95.3% vs 77.6%; p < 0.001) between the low- and high-risk groups. A significant improvement was confirmed using the prognostic methods for conventional TNM staging systems (p < 0.05).

CONCLUSIONS

The combination of SIRI with EBV DNA provided a more detailed understanding of patient risks, and enhanced risk discrimination in CALANPC.

Definitive radiation therapy and liver local therapy in de novo liver metastatic nasopharyngeal carcinoma: Large cohort study

BACKGROUND

We aimed to evaluate patients suitable for definitive radiation therapy (DRT) and liver local therapy (LLT) in addition to palliative chemotherapy (PCT) among those with de novo liver metastatic nasopharyngeal carcinoma (lmNPC).

METHODS

The overall survival (OS) and progression-free survival (PFS) rates were calculated and compared in 610 patients with lmNPC.

RESULTS

Both the PCT+DRT and PCT+DRT+LLT groups had better survival outcomes than the PCT group. Among patients with complete response/partial response (CR/PR) after PCT, no significant differences in survival rates were observed between those treated with PCT+DRT and PCT+DRT+LLT (2-year PFS: 27.0% vs. 32.9%, p = 0.263). Among patients with progressive disease/stable disease (PD/SD) after PCT, significantly better survival rates were observed in patients treated with PCT+DRT+LLT.

CONCLUSIONS

DRT might benefit patients with lmNPC regardless of the tumor response after PCT. For patients with CR/PR, LLT might not be needed. For patients with PD/SD, LLT might improve survival outcomes.

The prognostic role of plasma Epstein–Barr virus DNA levels in the middle of intensity-modulated radiotherapy to guide cisplatin dose recommendation in concurrent chemoradiation therapy in patients with locally advanced nasopharyngeal carcinoma: A large cohort study

Purpose

Methods and Materials

Results

Conclusions

First-Line Immunochemotherapy Versus Palliative Chemotherapy Plus Definitive Radiation Therapy for de novo Metastatic Nasopharyngeal Carcinoma: A Matched Cohort Study

BACKGROUND

The combined use of immune checkpoint inhibitors (ICIs) with palliative chemotherapy (PCT) is a promising first-line treatment for de novo metastatic nasopharyngeal carcinoma (mNPC). However, the efficacy of ICIs with PCT vs PCT with definitive radiation therapy (DRT) remain unclear.

METHODS

Patients with mNPC who received first-line immunochemotherapy (ICI + PCT) or PCT + DRT were included. Propensity score matching (PSM) was applied to balance potential confounders between patients who did and did not undergo DRT (at a ratio of 1:1). Progression free survival (PFS) and overall survival (OS) were compared between the 2 groups using a log-rank test and Cox proportional hazard model.

RESULTS

Among all participants, 149 received ICI + PCT. After PSM, 149 patients were included in the PCT + DRT group. First-line immunochemotherapy was associated with significantly improved PFS (median 9.0 months vs 12.0 months, P < .001) and OS (median 12.5 months vs 19.9 months, P < .001). Subgroup analysis revealed that tumor response to immunochemotherapy, metastatic organs, and number of metastatic sites potentially affected the efficacy of DRT after first-line immunochemotherapy.

CONCLUSION

Compared with PCT + DRT, first-line immunochemotherapy was associated with improved PFS and OS in patients with mNPC but not in patients with unfavorable tumor response and metastasis involving the liver, distant nodes, or multiple sites.

Advances in pathogenesis and precision medicine for nasopharyngeal carcinoma

Nasopharyngeal carcinoma (NPC) is a squamous carcinoma with apparent geographical and racial distribution, mostly prevalent in East and Southeast Asia, particularly concentrated in southern China. The epidemiological trend over the past decades has suggested a substantial reduction in the incidence rate and mortality rate due to NPC. These results may reflect changes in lifestyle and environment, and more importantly, a deeper comprehension of the pathogenic mechanism of NPC, leading to much progress in the preventing, screening, and treating for this cancer. Herein, we present the recent advances on the key signal pathways involved in pathogenesis of NPC, the mechanism of Epstein‐Barr virus (EBV) entry into the cell, and the progress of EBV vaccine and screening biomarkers. We will also discuss in depth the development of various therapeutic approaches including radiotherapy, chemotherapy, surgery, targeted therapy, and immunotherapy. These research advancements have led to a new era of precision medicine in NPC.

Management of first-line palliative chemotherapy for post-treatment metastasis after gemcitabine plus cisplatin induction chemotherapy: Gemcitabine plus cisplatin and non-gemcitabine plus cisplatin chemotherapy

BACKGROUND

To evaluate whether patients with post-treatment metastasis are suitable for GP first-line palliative chemotherapy (PCT) after undergoing GP IC.

METHODS

Seven hundred and forty-six patients with post-treatment metastasis after undergoing GP IC were eligible. Survival outcomes were compared.

RESULTS

Significant differences in survival rates were observed between patients treated with GP and non-GP chemotherapy (2-year progression-free survival [PFS]: 0.7% vs. 9.7%). We investigated survival outcomes of patients treated with GP PCT within 2 years after undergoing GP IC, treated with GP PCT 2 years after undergoing GP IC, and those of non-GP PCT patients (2-year PFS: 0.0%, 2.3%, 9.7%). However, there was no difference in the 2-year PFS between the patients that received GP PCT 2 years after undergoing GP IC and the non-GP PCT treated patients.

CONCLUSIONS

GP is not recommended for patients that have received GP IC within 2 years. Two years after undergoing GP IC, GP can be considered.

Prognostic Value of Plasma Epstein-Barr Virus DNA Levels Pre- and Post-Neoadjuvant Chemotherapy in Patients With Nasopharyngeal Carcinoma

Background

In this study, we evaluated the prognostic value of the plasma levels of Epstein-Barr virus (EBV) DNA in patients with nasopharyngeal carcinoma (NPC) at different treatment stages.

Methods

We retrospectively analyzed the Data of 206 patients with NPC. Pre-neoadjuvant chemotherapy (pre-NACT), post-NACT, post-radiotherapy, and post-treatment plasma EBV DNA levels were used to establish prognostic nomograms. The concordance index (C-index) and calibration curves were used to compare the prognostic accuracy of the nomograms. The results were confirmed in a validation cohort consisting of patients who were tested for EBV DNA levels at all four stages of treatment. The Kaplan-Meier method was used to calculate the progression-free survival (PFS) and overall survival (OS). Survival differences were calculated using the log-rank test.

Results

EBV DNA-positive patients had worse 3-year PFS and 5-year OS than EBV DNA-negative patients; this was true for pre-NACT (PFS: 82.7% vs. 57.3%, P < 0.001; OS: 90.9% vs. 68.7%, P = 0.08) and post-NACT (PFS: 85.0% vs. 50.6%, P < 0.001; OS: 91.7% vs. 65.7%; P = 0.001) EBV DNA levels but not for post-radiotherapy (PFS: 72.2% vs. 60.9%, P = 0.192; OS: 73.1% vs. 77.2%, P = 0.472) or post-treatment (PFS: 77.3% vs. 59.2%, P = 0.063; OS: 77.5% vs. 79.7%, P = 0.644) levels. Nomograms combining pre-NACT and post-NACT EBV DNA levels had a superior prognostic ability than those of post-radiotherapy and post-treatment EBV DNA levels.

Conclusion

Pre-NACT EBV DNA levels combined with post-NACT EBV DNA levels can more reliably predict survival outcomes in patients with NPC.

Residual Volume of Lymph Nodes During Chemoradiotherapy Based Nomogram to Predict Survival of Nasopharyngeal Carcinoma Patient Receiving Induction Chemotherapy

Purpose

To accurately stratify nasopharyngeal carcinoma (NPC) patients who were benefit from induction chemotherapy (IC) followed by chemoradiotherapy (CCRT), we established residual volume of lymph nodes during chemoradiotherapy based nomogram to predict survival for NPC patients.

Methods

Cox regression analysis were used to evaluate predictive effects of tumor volume parameters. Multivariate Cox regression analysis was used to identify the prognostic factors, and nomogram models were developed to predict survival of NPC patients receiving IC followed by CCRT.

Results

Compared with other tumor volumetric parameters, midRT GTVnd was the best predictive factor for OS (HR: 1.043, 95%CI: 1.031-1.055), PFS (HR: 1.040, 95%CI: 1.030- 1.051), and DMFS (HR: 1.046, 95%CI: 1.034 – 1.059) according to the HR of Cox regression analysis. Based on multivariate analysis, three nomograms included midRT GTVnd were constructed to predict 4-year survival. The C-index of nomograms for each survival endpoints were as follow (training cohort vs. validation cohort): 0.746 vs. 0.731 for OS; 0.747 vs. 0.735 for PFS; 0.768 vs. 0.729 for DMFS, respectively. AUC showed a good discriminative ability. Calibration curves demonstrated a consistence between actual results and predictions. Decision curve analysis (DCA) showed that the nomograms had better clinical predictive effects than current TNM staging system.

Conclusion

We identified the best volumetric indicator associated with prognosis was the residual volume of lymph nodes at the fourth week of chemoradiotherapy for patients receiving IC followed by CCRT. We developed and validated three nomograms to predict specific probability of 4-year OS, PFS and DMFS for NPC patient receiving IC followed by CCRT.

A systematic review and recommendations on the use of plasma EBV DNA for nasopharyngeal carcinoma

INTRODUCTION

Nasopharyngeal carcinoma (NPC) is an endemic malignancy in Southeast Asia, particularly Southern China. The classical non-keratinising cell type is almost unanimously associated with latent Epstein-Barr virus (EBV) infection. Circulating plasma EBV DNA can be a useful biomarker in various clinical aspects, but comprehensive recommendations and international guidelines are still lacking. We conducted a systematic review of all original articles on the clinical application of plasma EBV DNA for NPC; we further evaluated its strengths and limitations for consideration as standard recommendations.

METHODS

The search terms 'nasopharyngeal OR nasopharynx', and 'plasma EBV DNA OR cell-free EBV OR cfEBV' were used to identify full-length articles published up to December 2020 in the English literature. Three authors independently reviewed the article titles, removed duplicates and reviewed the remaining articles for eligibility.

RESULTS

A total of 81 articles met the eligibility criteria. Based on the levels of evidence and grades of recommendation assessed, it is worth considering the inclusion of plasma EBV DNA in screening, pre-treatment work-up for enhancing prognostication and tailoring of treatment strategy, monitoring during radical treatment, post-treatment surveillance for early detection of relapse, and monitoring during salvage treatment for recurrent or metastatic NPC. One major limitation is the methodology of measurement requiring harmonisation for consistent comparability.

CONCLUSIONS

The current comprehensive review supports the inclusion of plasma EBV DNA in international guidelines in the clinical aspects listed, but methodological issues must be resolved before global application.

An interpretable machine learning prognostic system for locoregionally advanced nasopharyngeal carcinoma based on tumor burden features

OBJECTIVES

We aimed to build a survival system by combining a highly-accurate machine learning (ML) model with explainable artificial intelligence (AI) techniques to predict distant metastasis in locoregionally advanced nasopharyngeal carcinoma (NPC) patients using magnetic resonance imaging (MRI)-based tumor burden features.

MATERIALS AND METHODS

1643 patients from three hospitals were enrolled according to set criteria. We employed ML to develop a survival model based on tumor burden signatures and all clinical factors. Shapley Additive exPlanations (SHAP) was utilized to explain prediction results and interpret the complex non-linear relationship among features and distant metastasis. We also constructed other models based on routinely used cancer stages, Epstein-Barr virus (EBV) DNA, or other clinical features for comparison. Concordance index (C-index), receiver operating curve (ROC) analysis and decision curve analysis (DCA) were executed to assess the effectiveness of the models.

RESULTS

Our proposed system consistently demonstrated promising performance across independent cohorts. The concordance indexes were 0.773, 0.766 and 0.760 in the training, internal validation and external validation sets. SHAP provided personalized protective and risk factors for each NPC patient and uncovered some novel non-linear relationships between features and distant metastasis. Furthermore, high-risk patients who received induction chemotherapy (ICT) and concurrent chemoradiotherapy (CCRT) had better 5-year distant metastasis-free survival (DMFS) than those who only received CCRT, whereas ICT + CCRT and CCRT had similar DMFS in low-risk patients.

CONCLUSIONS

The interpretable machine learning system demonstrated superior performance in predicting metastasis in locoregionally advanced NPC. High-risk patients might benefit from ICT.