Predicting and evaluating the efficacy of neoadjuvant therapy for rectal cancer are of clinical significance and health economic value. At present, exploring the methods of predicting and evaluating the efficacy of neoadjuvant therapy have become research hotspot, focus and difficulty at home and abroad. Radiomics and artificial intelligence (AI) are two rapidly developing technologies. It is worthy of integrating radiomics with AI to build a model for predicting and evaluating the efficacy of neoadjuvant therapy and support individualized clinical decision-making and treatment options. In this article, literature review related to neoadjuvant chemoradiotherapy for rectal cancer based on radiomics and AI was conducted, aiming to explore the prospect and advantages of radiomics and AI in the prediction and evaluation of neoadjuvant therapy.

Objective: To evaluate the safety and preliminary efficacy of total neoadjuvant therapy (TNT) in patients with locally advanced rectal cancer (LARC) with high risk factors. Methods: Data of 101 patients who were diagnosed with stage II-III rectal cancer with high risk factors and received TNT between March 2015 and January 2018 at West China Hospital of Sichuan University were analyzed retrospectively. Inclusion criteria: (1) patients were diagnosed with stage II-III rectal cancer by high-resolution MRI combined with CT and endorectal ultrasound; (2) at least one high risk factor: cT4a, cT4b, cN2, EMVI+, CRM+ and lateral lymph node+; (3) distance from tumor to anal verge was within 15 cm; (4) Eastern Collaborative Oncology Group (ECOG) performance status score was 0-1; bone marrow function, liver function and kidney function were suitable for chemoradiotherapy; (5) patients were treated with TNT strategy; (6) the follow-up data and postoperative pathological data were complete. Patients with previous rectal cancer surgery (except prophylactic colostomy), pelvic radiotherapy, and systemic chemotherapy, those with distant metastases, those without neoadjuvant radiotherapy, those receiving less than 4 cycles of neoadjuvant chemotherapy were excluded. The regimen of TNT: 3 cycles of induction CAPOX (oxaliplatin plus capecitabine) were followed by pelvic radiotherapy and concurrent CAPOX, then 3 cycles of consolidation CAPOX were delivered after radiotherapy. Total mesorectal resection (TME) or watch-and-wait strategy was selected according to the therapeutic effect and patients' wishes. Short-term efficacy, including tumor regression grade (TRG), pathological complete response (pCR), clinical complete response (cCR), postoperative complications within 30 days of surgery, and adverse events (AE) to radiotherapy and chemotherapy (measured using CTCAE 4.0) was analyzed. Results: The 101 patients included 68 males (67.3%) and 33 females (32.7%) with a median age of 54 years. The proportion of patients with cT4a, cT4b, cN2 and enlarged lateral lymph node was 13.9%, 29.7%, 56.4% and 43.6%, respectively. The mean cycle of neoadjuvant chemotherapy was 6.0±1.3. Seventy-five patients (74.3%) received at least 6 cycles of neoadjuvant chemotherapy and 100 (99.0%) completed radiotherapy. The mean cycle of induction and consolidation chemotherapy was 2.0±0.9 and 2.8±1.0 respectively. Most common grade 3 AE was leucopenia (n=13, 12.9%) and thrombocytopenia (n=7, 6.9%). Grade 3 diarrhea and radiation dermatitis were observed in 5 cases (5.0%) respectively. Grade 3 anemia and rectal pain were observed in 4 cases (4.0%) respectively. And rectal mucositis was observed in 2 cases (2.0%). Most of the AE was observed during concurrent chemoradiotherapy. No grade 4 or higher AE was observed. After TNT, 32 patients (31.7%) achieved pCR or cCR, and 62 patients (60.4%) achieved partial response (PR). Only 2 patients (2.0%) developed distant metastasis after chemoradiotherapy, while the other patients did not show disease progression. Seven patients (6.9%) with cCR refused surgery and selected watch-and-wait, while 7 patients without cCR still refused surgery. The other 87 patients (86.1%) underwent TME successfully. The mean interval from the completion of chemoradiotherapy to surgery was (20.1±8.5) weeks. The R0 resection rate was 97.7% (85/87).The morbidity of surgical complication was 16.1% (14/87), including pelvic infection or abscess in 6 cases (6.9%), anastomotic leakage in 3 (3.4%), hemorrhage in 2 (2.3%), and gastrointestinal dysfunction in 3 (3.4%). Pathological findings revealed that 24 cases (27.6%) had TRG 0, 20 (23.0%) had TRG 1, 30 (34.5%) TRG 2, and 13 (14.9%) TRG 3. Conclusion TNT is safe and has good short-term efficacy for locally advanced rectal cancer patients with high risk factors.