Prediction Role of Baseline Digital Substraction Angiography in GEP Neuroendocrine Liver Metastases Treated with TAE/ TACE

Xiaofen Li (  lxf0827@163.com ) West China Hospital of Medicine, Sichuan University West China Hospital https://orcid.org/0000-00021064-768X Linjuan Li West China Hospital of Medicine: Sichuan University West China Hospital Jin Pu West China Hospital of Medicine: Sichuan University West China Hospital Jiaojiao Suo West China Hospital of Medicine: Sichuan University West China Hospital Xuefeng Luo West China Hospital of Medicine: Sichuan University West China Hospital Xin You West China Hospital of Medicine: Sichuan University West China Hospital Zhengyin Liao West China Hospital of Medicine: Sichuan University West China Hospital Dan Cao West China Hospital of Medicine: Sichuan University West China Hospital https://orcid.org/0000-00032180-0464


Introduction
Gastroenteropancreatic neuroendocrine tumors (GEP-NETs) are rare and heterogeneous neoplasms originating from the diffuse neuroendocrine system of the gastrointestinal tract and pancreas. It is reported that the incidence rate of GEP-NETs has been rising in the last decades [1]. According to statistics from the Surveillance, Epidemiology, and End Results (SEER) program, the current annual age-adjusted incidence of GEP-NETs is estimated to be 3.56 per 100,000 persons in the United States [2]. Despite the indolent rate of tumor growth, over 20% NET patients present with distant metastasis at initial diagnosis, with liver the most common metastatic site [3].
According to current statistics, liver metastases occur in more than half of patients with GEP-NETs [4].
Current treatment approaches for management of liver metastatic GEP-NETs include surgery, somatostatin analogues, systemic chemotherapy, targeted drugs and intervention therapy. For patients with diffuse and bilobar liver metastases, transarterial embolization (TAE) or transarterial chemoembolization (TACE) is an important and appropriate treatment option to relieve symptoms and improve survival outcome, as suggested in several related guidelines [4,5].
Previous studies have suggested that patients with severe symptoms, lower liver tumor burden on imaging examinations or lower level of serum pancreastatin may potentially bene t from TAE/ TACE [6][7][8][9]. However, the prediction tool for therapeutic evaluation of TAE/ TACE is still unclear and optimal candidates for this procedure remain undetermined. The aim of this study was to assess the prediction role of baseline digital substraction angiography (DSA) in synchronous liver metastatic GEP-NETs patients who underwent TAE/ TACE procedures.

Patient Selection
The West China Hospital of Medicine, Sichuan University Ethic Committee for Clinical Investigation approved this study. And all methods were performed in compliance with the relevant guidelines and regulations. Medical information of patients diagnosed with synchronous liver metastatic GEP-NETs from January 1, 2016 to July 31, 2020 were reviewed retrospectively. Patients who met the following criteria were enrolled: 1) histologically con rmed GEP-NET; 2) clinically or pathologically diagnosed with synchronous liver metastasis between January 1, 2016 and July 31, 2020; 3) unresectable liver metastasis; 4) good performance status with an Eastern Cooperative Oncology Group (ECOG) score of < 2 (on a scale of 0 to 5, with a higher score indicating poorer performance status); 5) received TAE/ TACE procedure during disease process; 6) adequate clinical data.
Collected information included patient demographics, symptoms, hormonal functionality, tumor grade, TNM stage, baseline DSA ratio, baseline computed tomography (CT) ratio, liver speci c tumor response, systemic therapy, and serum tumor marker (CEA and NSE) levels. In this study, liver speci c tumor response was de ned as radiographic response of liver metastases to treatment according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 [10]. Baseline DSA ratio was de ned as the density of the largest target lesion on DSA imaging (houns eld unit, HU)/ the density of liver background (HU) before TAE/ TACE, and CT ratio meant the baseline density of the largest target lesion on CT imaging HU / the density of abdominal aorta (HU) during arterial phase. When the tumor grades between primary and metastatic tumors were different, the higher grade were taken and recorded.

Statistical Analysis
Quantitative data were compared by Student's t test and categorical data by Chi-square test. Statistical tests were two sided and P < 0.05 was considered statistically signi cant. All analyses were performed using SPSS Statistics 22.0 (IBM, Armonk, NY, USA).

Patient characteristics
In total, 22 patients with unresectable liver metastatic GEP-NETs were enrolled. Table 1 listed patents' clinical and pathological characteristics. The mean age at diagnosis was 49.3±9.8 years (95% con dence interval (CI): 44.9-53.6 years). More than half of the patients (68.2%) were females. All of the patients were in good performance status with ECOG scores 0~1. The vast majority of primary tumors were located in pancreas (86.4%, 19/22) and non-functional (91.0%, 20/22). Most of patients (81.8%) presented with G2, while 13.6% with G3 and 4.5% with G1. The mean largest diameter of liver metastases before TAE/ TACE was 6.3±4.7 cm (95% CI: 3.8-7.5 cm). Most patients presented with elevated serum NSE (59.1%) and normal serum CEA (86.4%) before TACE. All patients underwent TAE/ TACE procedure after assessment of oncologists and interventional radiologists. During TACE procedure, with the help of DSA examination, hepatic artery was selected using a 5 French catheter through femoral arterial access. Next, tumor-feeding vessels were super-selected and infused with a mixture of iodized oil (10-20 mL) and cytotoxic agents ( uorouracil 1000 mg and doxorubicin 50 mg). Then embolization was performed using gelatin sponge particles to complete blood ow stagnation. TAE procedure was similar to TACE, but without administration of iodized oil or cytotoxic agents, and completed by occlusion of the hepatic artery branch with polyvinyl alcohol particles.
As systemic therapy, 16 (72.7%) patients received somatostatin analogues, 5 patients were prescribed with targeted agents, and 6 patients received chemotherapy.  Baseline Dsa Prediction Role In Therapeutic Evaluation Therapeutic response was evaluated by abdominal computerized tomography (CT) or magnetic resonance imaging (MRI) at 2-to 3-month intervals according to RECIST 1.1 criteria. The largest target liver lesion of every patient was selected for therapeutic evaluation. The liver speci c overall response rate (ORR) was 45.5% (10/22), with 10 patients evaluated as partial response (PR) and 0 complete response (CR). Figure 1 showed the best diameter change of target liver lesions from baseline. Among all the patients, the liver speci c median progression free survival (PFS) was 17.5 months (95% CI 4.8-30.2 months), and overall survival (OS) were not reached ( Figure 2). Kaplan-Meier analysis showed that patients with lower hepatic tumor burden (<50%) had signi cant longer PFS than higher hepatic tumor burden (≥50%) (62.2 vs 13.3months, P=0.030, Figure 2c). Other groups strati ed by CT ratio, DSA ratio, age, sex, or NSE level showed no signi cant differences in PFS (Figure 2), which might be partly attributed to small sample size. Multivariate COX regression analysis revealed that hepatic tumor burden was an independent prognostic factor for PFS (Table 3). The average baseline DSA ratio of the largest target lesion (the density of the target lesion on DSA imaging / the density of liver background before TAE/ TACE) was 0.64±0.15. And the average baseline DSA ratio in responsive group was signi cantly lower compared with that in nonresponsive group (0.57±0.13 versus 0.70±0.15, P=0.037, Figure 3a). Furthermore, patients with a DSA ratio ≤0.64 were more responsive to TAE/ TACE than those with a DSA ratio 0.64 (58.3% versus 30%), but the P value was greater than 0.05 probably due to small sample size. Figure 3b showed therapeutic evaluation in patients with a low DSA ratio (≤0.64) and a high DSA ratio ( 0.64). Figure 4a~c and 4d~f were representative images of patients with high and low DSA ratios. Meanwhile, we also assessed the role of baseline CT ratio of the largest target lesion (the density of the target lesion on CT imaging / the density of abdominal aorta during arterial phase) in therapeutic evaluation. The results showed that among all the patients, the average baseline CT ratio was 0.34±0.09. The CT ratios between responsive group and nonresponsive group were not statistically different (0.30±0.06 versus 0.36±0.11, P=0.149, Figure 3b).

Discussion
In this study, we summarized the characteristics and therapeutic response of patients with unresectable liver metastatic GEP-NETs treated with TAE/ TACE. We found that baseline DSA ratio of the target lesion could predict treatment response usefully. And patients with lower hepatic tumor burden might indicate better prognosis. To the best of our knowledge, this is the rst study investigating the association between baseline DSA features and therapeutic response to TAE/ TACE.
Several studies investigated characteristics and survival outcome of patients with neuroendocrine liver metastases treated with TAE/ TACE[6-9, 11-13]. The hepatic ORR after TAE/ TACE procedure was reported to be 20-40%, which was in accordance in our study. In the study conducted by Dhir M. et al, age, grade, liver tumor burden, and serum chromogranin A level were independent prognostic factors for overall survival in neuroendocrine tumor liver metastases received TACE [11].  [15]. But this method greatly depended on the professional level of radiologists and consequently had low consistency. In our study, we used a quantitative data (DSA ratio) to measure tumor vascularity and evaluate response, which was simple and accurately measurable. Even doctors not majored in radiology could easily learn and master this new method.
Some limitations exist in our study. First, selection bias is inevitable due to the retrospective nature. Second, for the rarity of liver metastatic GEP-NETs, the sample size of our study is very small. We are making our effort to collecting more cases in the future.

Conclusions
In conclusion, baseline DSA ratio of the largest target lesion is a simple and potentially useful approach to predict therapeutic effect of TAE/ TACE in liver metastases from GEP-NETs. In addition, patients with lower hepatic tumor burden might indicate better prognosis. Large prospective study is warranted to con rm this.  Figure 1 The best diameter change of target liver lesions from baseline (overall response rate=45.5%).