Qiyue Chen,Zhongliang Ning,Zhiyu Liu,Yanbing Zhou,Qingliang He,Yantao Tian,Hankun Hao,Wei Lin,Lixin Jiang,Gang Zhao,Ping Li,Chaohui Zheng,Changming Huang,on behalf of the Study Group for Gastric Neuroendocrine Tumors

1Department of Gastric Surgery,Fujian Medical University Union Hospital,Fuzhou 350001,China;2 Department of General Surgery,Fujian Medical University Union Hospital,Fuzhou 350001,China;3Key Laboratory of Ministry of Education of Gastrointestinal Cancer,Fujian Medical University,Fuzhou 350004,China;4 Fujian Key Laboratory of Tumor Microbiology,Fujian Medical University,Fuzhou 350004,China;5Department of Gastrointestinal Surgery,West District of the First Affiliated Hospital of USTC,Division of Life Sciences and Medicine,University of Science and Technology of China,Hefei 230001,China;6Department of Gastrointestinal Surgery,Affiliated Hospital of Qingdao University,Qingdao 266000,China;7Department of Gastrointestinal Surgery,the First Affiliated Hospital of Fujian Medical University,Fuzhou 350005,China;8Department of Pancreatic and Gastric Surgery,National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College,Beijing 100021,China;9 Department of General Surgery,Huashan Hospital,Fudan University,Shanghai 200040,China;10Department of Gastrointestinal Surgery and Gastrointestinal Surgery Research Institute,the Affiliated Hospital of Putian University,Putian 351106,China;11 Department of Gastrointestinal Surgery,Yantai Yuhuangding Hospital,Yantai 264099,China;12Department of Gastrointestinal Surgery,Renji Hospital,Shanghai Jiaotong University,Shanghai 200127,China

Abstract Objective:Quality assurance is crucial for oncological surgical treatment assessment.For rare diseases,singlequality indicators are not enough.We aim to develop a comprehensive and reproducible measurement,called the“Textbook Outcome”(TO),to assess the quality of surgical treatment and prognosis of gastric neuroendocrine carcinoma (G-NEC) patients.Methods:Data from patients with primary diagnosed G-NEC included in 24 high-volume Chinese hospitals from October 2005 to September 2018 were analyzed.TO included receiving a curative resection,≥15 lymph nodes examined,no severe postoperative complications,hospital stay ≤21 d,and no hospital readmission ≤30 d after discharge.Hospital variation in TO was analyzed using a case mix-adjusted funnel plot.Prognostic factors of survival and risk factors for non-Textbook Outcome (non-TO) were analyzed using Cox and logistic models,respectively.Results:TO was achieved in 56.6% of 860 G-NEC patients.TO patients had better overall survival (OS),disease-free survival (DFS),and recurrence-free survival (RFS) than non-TO patients (P＜0.05).Moreover,TO patients accounted for 60.3% of patients without recurrence.Multivariate Cox analysis revealed non-TO as an independent risk factor for OS,DFS,and RFS of G-NEC patients (P＜0.05).Increasing TO rates were associated with improved OS for G-NEC patients,but not hospital volume.Multivariate logistic regression revealed that nonlower tumors,open surgery,and ＞200 mL blood loss were independent risk factors for non-TO patients (P＜0.05).Conclusions:TO is strongly associated with multicenter surgical quality and prognosis for G-NEC patients.Factors predicting non-TO are identified,which may help guide strategies to optimize G-NEC outcomes.

Keywords:Textbook Outcome;gastric neuroendocrine carcinoma;surgical quality;prognosis;risk factor

Introduction

Gastric cancer (GC) is the second most common malignant tumor and the third cause of cancer-related deaths (1,2).However,gastric neuroendocrine carcinoma (G-NEC),a particular type of GC,has very rare morbidity,despite having a common origin with GC.In recent years,as surgical knowledge of G-NEC has advanced,diagnosis and treatment for this type of tumor has also become increasingly standardized. Nevertheless,a G-NEC prognosis still leads to more adverse outcomes than common GC (3-5).Currently,many researchers are exploring clinical factors and novel therapies related to GNEC to improve prognosis.In addition to the tumor’s effect itself,it is being gradually recognized that quality control of perioperative health services also has a profound effect on the patient’s survival.For more common types of tumors,such as GC,some major clinical centers may deal with numerous cases in a short time (6).Currently,the largest number of GC cases reported was 10,000 in a single hospital (7).Consequently,even for a single center,seeking appropriate medical quality control targets can be easily achieved by gradually accumulating clinical experience and sequentially increasing the recognition of pathological features.Surgeons often use a single stable parameter to complete the quality control of surgery amongst a large amount of accumulated treatment data.However,similar to other rare tumors,G-NEC cases often occur with a scatterpoint distribution in time and place,making long-term follow-ups difficult.Simultaneously,international reports are generally small sample studies (8,9),and it is difficult to summarize effective,reproducible quality control metrics with a single-center study.Furthermore,an excessive focus on single parameters may disregard the multidimensionality of the complex surgical therapy of rare diseases.Therefore,it is necessary to carry out multicenter studies to integrate the available scattered data and find an ideal composite index for surgical quality assessment.In 2013,Kolfschotenet al.(10) first proposed the composite index of“Textbook Outcome (TO)”,whose value has been gradually recognized in the assessment of care quality and prognosis of common tumors.Given its combination of available short-term data,including radical resection,number of lymph node dissections,postoperative complications,postoperative hospital stay,and postoperative readmission,TO is technically more suitable as an index of surgical quality control for rare diseases.However,there is still no literature reporting on the potential value of TO for the treatment of G-NEC patients with gastrectomy,and their long-term prognostic outcome.Hence,this study aimed to collect the clinical data of GNEC patients from 24 Chinese hospitals,to carry out the first multicenter study revealing the value of TO for this rare disease.We developed a comprehensive and reliable indicator for those who received radical G-NEC surgery and analyzed the potential of TO in assessing surgical quality and prognosis.We hope to provide a tool to evaluate surgical quality systematically,improve the standards of surgical treatment,and guide clinical practice about the best care resources.

Materials and methods

Study design and cohort

We retrospectively analyzed the data of 860 G-NEC patients with complete information between October 2005 and September 2018,from the Study Group for Gastric Neuroendocrine Tumors (SGGNET) (11,12).SGGNET included 24 high-volume Chinese hospitals.The following exclusion criteria were applied:1) missing survival data;2)Mx and preoperative or postoperative M1 diagnosis;or 3)patients without surgery;After these evaluations,860 GNEC patients were considered.The selection scheme is provided inSupplementary Figure S1.

This study was conducted with the approval of the Institutional Review Boards of all participating hospitals.All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions.Informed consent or a substitute for it was obtained from all patients for inclusion in the study.

Definitions

In this study,TO refers to a combination of five quality metrics related to oncologic resection,postoperative course,and discharge of patients undergoing G-NEC surgery.These include:a curative resection;≥15 lymph nodes (LNs) examined; no severe postoperative complications;a hospital stay ≤21 d;and no hospital readmission ≤30 d after discharge.When all parameters were met,TO was achieved.There is no generally acknowledged definition of TO for G-NEC.With respect to the previously published literature (10,13-20),we have modified the definition of TO to conform to this analysis and the data available from SGGNET.Our TO,which consists of five indicators,was collected through the hospital elective system.Postoperative complications of grade II or higher,according to the Clavien-Dindo classification,were considered severe.

At present,there is no obvious distinction between surgical treatment strategies for mixed neuroendocrineeuroneuroendocrine neoplasms and neuroendocrine carcinoma (NEC) (21,22);hence,we combined them in the analysis.The staging was derived from the 8th American Joint Commission on Cancer (AJCC) criteria.According to the 2019 World Health Organization (WHO) classification(23,24),patients were sorted as G1 (Ki67≤2%),G2(2%＜Ki67≤20%),or G3 (Ki67＞20%).

Overall survival (OS) was calculated from the date of surgery to death from any cause.Recurrence-free survival(RFS) was calculated from the date of surgery to locoregional or distant recurrence due to any cause.Recurrence was defined by imaging evaluation,cytology,or tissue biopsy in combination with the clinical history and physical examination at the earliest date.Recurrence was diagnosed based on radiologic findings or biopsies of suspicious lesions,when possible.Disease-free survival (DFS) was calculated from the date of surgery to any loco-regional or distant recurrence,other primary tumors,or death from any cause.Postoperative follow-ups were performed every 3-6 months for the first 3 years,and then every 6-12 months from years 3 to 10.Most routine patient follow-up appointments included a physical examination,laboratory tests,chest radiography,abdominal ultrasonography or computed tomography (CT),and an annual endoscopic examination.The median follow-up time in this study was 55 months.The maximum follow-up date was October 2018.The medium follow-up time was 55 months.

Statistical analysis

A Student’sttest or Mann-Whitney U test was used for continuous variables.We used a χ2test or Fisher’s exact test to compare the categorical variables of the clinical characteristics. First,we calculated the number and proportion of patients for whom each outcome indicator was realized. Then we calculated the number and proportion of patients for whom each consecutive outcome indicator was realized (namely,those who simultaneously met the previous and subsequent metric criteria) (10).We estimated the median survival using the Kaplan-Meier method.The survival rates were then compared using a log-rank test.Landmark analysis for G-NEC patients was used to eliminate the immortal bias.The landmark time point was at day 31 of follow-up.The association of relevant clinicopathologic variables with OS,DFS,and RFS was assessed using a Cox proportional hazards model.We assessed the risk factors of non-Textbook Outcome(non-TO) using multivariate logistic regression analysis,incorporating other potential explanatory variables.Stepwise backward variable removal was applied to the multivariate model,to identify the most accurate set of predictors (25).The TO rate was considered a categorical variable.We calculated the TO rate among different hospitals and ranked these hospitals accordingly.The cutoff values were lower quartile-50% and upper quartile-60%.The centers were divided into three groups (TO≤50% group,50%＜TO≤60% group,and TO＞60% group)to further analyze the relationship between TO and 5-year OS.Hospital volume was also considered a categorical variable and used ranking hospitals in order to increase total volume and select cut-off points in the interquartile range.Specifically,we applied the following distinction:low-volume (n≤50 patients),medium-volume (50 patients＜n≤100 patients),and high-volume (n＞100 patients)centers. Then we calculated three levels for the corresponding TO rates,and investigated the relationship among hospital volume,TO rate,and OS (15,26-28).

To analyze hospital variation in TO,case mix-adjusted hospital results were calculated.The possible associations between patient characteristics and TO were analyzed to subsequently adjust hospital TO rates for the case-mix factors.Therefore,patient and tumor characteristics were entered in a multivariate logistic regression model at a Pvalue of 0.05,using an ENTER model.To adjust for casemix factors,several variables were analyzed including age,American Society of Anesthesiologists (ASA) score,surgical type,blood loss,and tumor location.Individual hospital results were displayed using funnel plots,combining scatter plots and a sequence of 95.0% and 99.8% confidence intervals (95.0% CIs and 99.8% CIs) (29).To minimize statistical artefacts resulting from the small sample size,hospitals with ＜11 G-NEC resections were excluded from this analysis.We also performed risk-adjusted funnel plots of all hospitals to show the details of hospitals with ＜11 GNEC resections.

Sankey plots were applied to analyze the relationship between different TO items and survival outcome.The survival status within Sankey plots was recorded until follow-up was completed.Statistical analyses were performed using IBM SPSS (Version 20.0;IBM Corp.,New York,USA) and R software (Version 3.5.1;R Foundation for Statistical Computing,Vienna,Austria).P＜0.05 were considered statistically significant.

Results

General clinical and pathological data

Baseline characteristics for the TO and non-TO groups are presented inSupplementary Table S1.Altogether,860 patients were included in the study;they were divided into a TO group (487 patients,56.6%) and a non-TO group(373 patients,43.4%).There were significant differences in period,age,ASA score,surgical type,blood loss,and tumor location (P＜0.05).However,no significant differences in sex,body mass index (BMI),pT and pN stage,grade,tumor size,neoadjuvant chemotherapy,and chemotherapy were observed (P＞0.05).In total,23 (6.2%) patients did not achieve the R0 margin in the non-TO group.

TO after resection of G-NEC

In accordance with the methods,we individually calculated each quality metric:97.3% (837/860 patients) achieved curative resection,79.7% (685/860 patients) had ≥15 LNs examined,80.6% (693/860 patients) had no severe postoperative complications,87.1% (749/860 patients) had a hospital stay ≤21 d,and 98.6% (848/860 patients) had no hospital readmission ≤30 d after discharge.The quality metrics that had the most negative impact on the proportion achieving TO were ≥15 LNs examined and no severe postoperative complications,which occurred in only 79.7% and 80.6% of cases,respectively.

We then sequentially calculated the five quality metrics:curative resection was achieved in 97.3% of patients;curative resection and ≥15 LNs examined were achieved in 78.6%;curative resection,≥15 LNs examined,and no severe postoperative complications was achieved in 62.9%;curative resection,≥15 LNs examined,no severe postoperative complications,and postoperative hospital stay ≤21 d was achieved in 57.2%.Approximately,56.6% of patients met all five TO quality metrics.Hence,487 patients achieved TO (Figure 1).

Supplementary Figure S2shows the changing trends of TO during different periods from 2005 to 2018.Except for a slightly lower TO rate from 2011-2013 than from 2008-2010,the TO rate gradually increased over time,with 2005-2007 showing the lowest TO rate (only 40%),and 2017-2018 showing the highest (72.7%).

Figure 1 Textbook Outcome:a composite measure of outcome parameters in patients undergoing surgery for G-NEC.G-NEC,gastric neuroendocrine carcinoma.

Survival analysis

Survival curves revealed that OS,DFS,and RFS of TO patients were significantly better than those of non-TO patients (all P＜0.05;Figure 2).Landmark analysis also showed that results of OS,DFS,and RFS of TO patients were consistent with Kaplan-Meier survival curves.Landmark analysis revealed that at day 31 of follow-up,OS,DFS,and RFS of TO patients were similar to those of non-TO patients (P＞0.05).After the 31-day follow-up,OS,DFS,and RFS of TO patients were significantly better than those of non-TO patients (P＜0.05) (Supplementary Figure S3).Because TO consists of five quality metrics,through the dynamic display function of the Sankey plot,the relationship between the five metrics and the individual’s final survival status was displayed.

InFigure 3,G-NEC patients were gradually isolated when challenging each TO item individually.The prognostic outcomes of“Alive”mostly consisted of TO patients (62.2%) who definitely met the five metrics,and only the rest of“Alive”outcome (37.8%) were from non-TO patients (Figure 3A).Similarly,among the analysis of recurrence outcomes,the prognostic outcomes with“no recurrence”were mostly made up of TO patients (60.3%),whereas non-TO patients accounted for only 39.7% and presented no recurrence (Figure 3B).In this process,every additional unmet metric will be prone to a more dismal long-term outcome for the patients.

Considering the independent prognostic factors affecting prognosis (Table 1),Cox univariate analysis showed that TO,hospital volume,age,ASA score,tumor location and size,pT and pN stages,and G grade were factors significantly affecting the patients’ OS (P＜0.05).Further multivariate analyses revealed that overlapping location,higher pT stage,lymph node metastasis,and non-TO were independent risk factors for OS,but the medium-volume center (50＜n≤100) was independent protective factors(P＜0.05).In parallel,in G-NEC patients,non-TO was also an independent risk factor for both DFS and RFS (P＜0.05)(Supplementary Table S2,S3).

Table 1 Univariate and multivariate Cox analysis of clinicopathological factors for OS

Table 1 (continued)

Hospital volume,TO,and OS

Usually,single-center hospital volume is an important index for surgical quality assessment.However,for GNEC,our multicenter study showed that high-volume centers did not stably reflect better long-term survival(Figure 4A).Even medium-volume centers presented a better OS than high-volume ones (P=0.036).However,by observing the relationship between TO and survival,it can be seen that as TO gradually increases,the long-term survival rate of patients can gradually increase (Figure 4B,C).As shown inSupplementary Figure S4,at day 31 of follow-up,there was no significant difference in OS among different volume hospitals and different levels of TO rates,respectively (P=0.210,P=0.312).After the 31-day followup,high-volume centers did not stably reflect better longterm survival,while medium-volume centers even presented a better OS.Hence,by comparing with TO,we found that the long-term survival of G-NEC patients improved with increasing TO rates.The three-dimensional stereoscopic diagram and funnel plot showed the relationship among different volumes,TO,and survival.When the hospital volume increased,the 5-year survival of G-NEC patients did not improve,and the hospital volume level was not associated with 5-year survival (Pearson test,P=0.567;Figure 5A).However,with increasing TO rates,the 5-year survival of G-NEC patients improved (Pearson test,P=0.013).

Figure 2 Prognosis of TO patients was superior to that of non-TO patients.(A) OS (P=0.003);(B) DFS (P=0.004);(C) RFS (P=0.005).TO,Textbook Outcome;OS,overall survival;DFS,disease-free survival;RFS,recurrence-free survival.

Figure 3 Sankey plot dynamically displays flow relationship between TO and its five different metrics,between TO and the final prognosis outcome.(A) TO flows to the survival outcome;(B) TO flows to the recurrence outcome.TO,Textbook Outcome;OS,overall survival.

TO rates for individual centers were adjusted by casemix factors (Figure 5B).The adjusted TO rates ranged from 9.1% to 86.5% for G-NEC patients.With increasing numbers of cases,the TO rates did not significantly improve,suggesting that the realization rate of TO was not directly associated with the increasing hospital volume.Supplementary Figure S5shows the risk-adjusted performances of all hospitals.Supplementary Table S4shows the clinicopathologic description between medium-and high-volume centers.

Figure S1 Patient flow diagram.NET,neuroendocrine tumor.

Figure S2 Trends of TO rates from 2005 to 2018.TO,Textbook Outcome.

Figure S3 Landmark analysis of discriminating between events occurring before and after 31 d of follow-up.(A) OS (Pbeforethelandmarktimepoint=0.275,Pafterthelandmarktimepoint=0.004);(B) DFS (Pbeforethelandmarktimepoint=0.160,Pafterthelandmarktimepoint=0.015);(C) RFS (Pbeforethelandmarktimepoint=0.130,Pafterthelandmarktimepoint=0.011).OS,overall survival;DFS,disease-free survival;RFS,recurrence-free survival.

Figure S4 Landmark analysis among different hospital volumes,TO,and OS.(A) OS of patients within different hospital volumes (Pbeforethelandmarktimepoint=0.210,Pafterthelandmarktimepoint＜0.001,Pn≤50 vs.50＜n≤100=0.038;Pn≤50 vs.n＞100＜0.001;P50＜n≤100 vs.n＞100=0.040);(B) OS of patients within different TO rates (Pbeforethelandmarktimepoint=0.312,Pafterthelandmarktimepoint＜0.001,PTO≤50% vs.50%＜TO≤60%=0.033;PTO≤50% vs.TO＞60%＜0.001;P50%＜TO≤60% vs.TO＞60%=0.025).TO,Textbook Outcome;OS,overall survival.

Figure 4 Relationship among different hospital volumes,TO,and OS.(A) OS of patients within different hospital volumes (Pn≤50 vs.50＜n≤100=0.036;P50＜n vs.n＞100＜0.001;P50＜n≤100 vs.n＞100=0.038);(B) OS of patients within different TO rates (PTO≤50% vs.50%＜TO≤60%=0.030;PTO≤50%vs.TO＞60%＜0.001;P50%＜TO≤60% vs.TO＞60%=0.023);(C) 5-year OS of patients within different TO rates (P=0.006).TO,Textbook Outcome;OS,overall survival.

Figure 5 Relationship among hospital volumes,TO rate,and 5-year OS.(A) Three-dimensional stereoscopic diagram.The red dots represent the projection from the hospital volume (“n”,in the XZ plane;coordinate system of the hospital volume and 5-year OS).OS within the three levels was not improved with increasing levels of hospital volume.Pearson test revealed no association between hospital volume and 5-year OS (P=0.567).The blue dots represent the projection from different TO rates (“n”,in the YZ plane;coordinate system of the TO rate and 5-year OS).OS within three groups was improved with increasing TO rate.Pearson test revealed a significant association between TO rate and 5-year OS (P=0.013);(B) Hospital variation in risk-adjusted percentages of TO.The gray dots indicate individual institutions,the green solid lines and red dotted lines indicate the 95% and 99.8% CIs,respectively.To minimize statistical artefacts because of the small sample size,hospitals with ＜11 G-NEC resections were excluded from this analysis.TO,Textbook Outcome;OS,overall survival;95% and 99.8% CIs,95% and 99.8% confidence intervals;G-NEC,gastric neuroendocrine carcinoma.

Figure S5 Hospital variation in risk-adjusted percentages of TO.Black dots indicate individual institutions and green solid lines and red dotted lines indicate 95% and 99.8% CIs,respectively.TO,Textbook Outcome;G-NEC,gastric neuroendocrine carcinoma;95% and 99.8% CIs,95% and 99.8% confidence intervals.

TO-associated factors

To explore whether non-TO can be predicted preoperatively and intraoperatively,the logistic univariate analysis of risk factors for non-TO patients is shown inTable 2.According to this,age,ASA score,tumor location,surgical type,and blood loss were significant factors resulting in non-TO outcomes (P＜0.05). Further multivariate analysis showed that non-lower tumors,open surgery,and blood loss ＞200 mL were independent risk factors for non-TO patients (P＜0.05).The analysis of the risk factors affecting TO patients is shown inSupplementary Table S5.When the number of risk factors was 0,1,2,and 3,the proportion of TO patients was 75.4% (52/69),63.9%(212/332),50.8% (197/388),and 36.6% (26/71),respectively.With the accumulation of risk factors,the proportion of patients achieving TO significantly decreased (P＜0.001).

Table S1 Clinicopathologic description of all G-NEC patients (N=860)

Table S1 (continued)

Table S2 Univariate and multivariate Cox analysis of clinicopathological factors for DFS

Table S2 (continued)

Table 2 (continued)

Table S3 Univariate and multivariate Cox analysis of the clinicopathological factors for RFS

Table S3 (continued)

Table S4 Clinicopathologic description of G-NECs between medium-and high-volume centers

Table S5 Risk factors related to TO

Discussion

Although information on individual outcome indicators may be useful in evaluating medical quality improvement,it is difficult to compare hospital performance through individual indicators as a hospital may perform better in some than in others,since these parameters are often not related (10,16,30).The medical service level cannot be scientifically reflected by only one point;rather,it should be indicated by a comprehensive group of items.The evaluation of surgical treatments should be multidimensional.The Society for Thoracic Surgeons was one of the first to start a clinical audit to monitor their results (31,32).Then,O’Brienet al.developed and analyzed a method of composite scoring for cardiac surgery,which was described as the“all-or-none”method,and represented the base for our TO (33).This provided a feasible and informative indicator,suitable for comparison in multicenter studies.This work demonstrated the effectiveness of a composite indicator,consisting of five single parameters,in the assessment of medical treatments.Among G-NEC patients,56.6% achieved TO.After adjusting for the case mix,we found that there were differences in TO rates among different centers,where the top rate reached 86.5%,but the lowest rate was only 9.1%.Funnel plots indicated that TO rates were not enhanced with increasing hospital volume,implying no direct association between them.When analyzing the relationship between hospital volume and survival,the prognosis within medium-volume centers was better than within low-and high-volume centers. Based on previous experience,increasing numbers of surgical cases are often accompanied by enhanced surgery quality,thereby improving prognosis for patients.Due to the specificity of G-NEC,the prognosis for such patients did not significantly improve with hospital volume,which is significantly different compared with common GC (26,27,34,35).For patients undergoing curative G-NEC surgery,the surgical process can be considered safe if no adverse outcomes (severe postoperative complications,prolonged hospital stay,and readmission) have occurred,and effective if complete tumor removal and adequate lymphadenectomy have been achieved.These parameters have been included in the definition of TO.Our results showed that failure in retrieving at least 15 LNs and severe postoperative complications had the greatest negative impact on TO for G-NEC patients.Therefore,for clinical surgeons,it will be paramount to improve the level of intraoperative LN dissection and reduce the incidence of severe postoperative complications.Simultaneously,a better understanding of different factors that lead to TO success or failure may potentially help health service providers to further improve postoperative management and care quality,as well as to reduce hospital costs (36,37).Altogether,TO provides caregivers with important information on patient therapeutic feedback,and this may drive quality improvements in hospitals.For patients,TO also discloses more information and available choices for favorable outcomes,thereby selecting specific care-service resources.

For hospitals,TO provides information on how often treatment is successful and may drive quality improvement.TO may be useful in selective contracting as it summarizes indicators of patient safety,effectiveness,and efficiency.

NEC is a higher malignant degree type of gastric neuroendocrine neoplasm,whose incidence and morbidity increase every year (38,39).Currently,among the prognostic factors associated with G-NEC,the most recognized index includes Ki67 and a constantly updated TNM stage (4,5).However,the above information is generally retrieved from tumor pathology and anatomy.Although many studies reported the effects of surgery onthe prognosis of NEC,the combination of surgical quality and postoperative management,namely,the impact of TO on long-term prognosis,has not been evaluated.Our results showed that TO was an independent risk factor for G-NEC patients.Survival analysis showed that the prognosis of TO patients was significantly better than that of non-TO patients.Furthermore,an escalating TO rate was strongly associated with improved survival,implying that successful operation and careful postoperative management were related to the long-term outcomes of patients.In the study,factors related to tumor and surgery affected the likelihood of achieving TO after G-NEC gastrectomy.More precisely,risk-related factors that may increase the probability of non-TO included non-lower tumors,open surgery,and ＞200 mL blood loss.Tumors located in the non-lower part of the stomach may implicate a wider range gastrectomy,such as a total gastrectomy,with a possible upgrade of the operation difficulty.Compared with minimally invasive procedures,such as laparoscopy or robotic surgery,open surgery implies a larger surgical incision,leading to greater damage,which would be detrimental to postoperative early recovery.Finally,excessive blood loss during surgery will increase the postoperative recovery time,reducing the probability of achieving TO.A clinical surgeon should identify patients with high-risk factors and develop more detailed treatment options for them.TO encloses the acknowledged prognostic factors,including radical resection,adequate LN dissection,and no severe postoperative complications.The prognostic value of the dissection of at least 15 LNs for GC patients has already been proven by many researchers (40-44).Severe postoperative complications may also play a role in long-term survival.Anastomotic leakage and infectious complications are associated with disease recurrence in various tumor types (45-49).

To our knowledge,this is the first study on a composite evaluation measure for surgical quality and comprehensive predictor for long-term outcome in rare cancers.It included a large amount of data and long-term follow-up.However,some limitations exist.Firstly,other data,such as those on targeting and endocrine treatment were incomplete,which may have inevitably caused some bias.Furthermore,there is no acknowledged definition of TO;thus,selecting different measures of surgical quality indicators may lead to different outcomes.Lastly,because of their composition,the five TO metrics could possibly be refined by adding weights to different outcome measures,making some outcomes more important than others.However,no clear evidence or resource from which to derive this weight is available.Therefore,any weights added to the TO metrics would be subjective and diminish its simplicity and usefulness (10).

Moreover,this study only included five items of TO,yet the multi-center results demonstrated the values of TO in the surgical quality control and long-term prognosis for GNEC.Such a complex indicator TO was characteristic of easy promotion and high reproducibility in the real world.It is worth noting that TO is not designed to replace the individual quality indicators,but is meant as an addition to them.Considering medical and surgical complexity and specific differences among individual patients,our results do not imply that patients who did not meet all TO indicators received incorrect treatment.In the future,we look forward to exploring more clinical studies on the surgical treatment of GC and related rare diseases by combining hospitals with different volumes in China to validate our findings.

Conclusions

Hospital volume was not a good indicator for assessing surgical quality of G-NEC.TO is strongly associated with multicenter surgical quality and prognosis for G-NEC patients.Factors predicting non-TO are identified,which may help guide strategies to optimize G-NEC outcomes.

Acknowledgements

This study was supported by scientific and technological innovation joint capital projects of Fujian province (No.2018Y9041);National Natural Science Foundation of China (No.82002462);China Scholarship Council (No.201908350095);Provincial Natural Science Foundation Project (No.2020J011001) and Fujian Medical University Outstanding Young Cultivation Project (No.2020PYY002).

Footnote

Conflicts of Interest:The authors have no conflicts of interest to declare.