Technique Modification For Laparoscopic Lateral Lymph Node Dissection In Rectal Cancer,Node-by-Node Matching And Radiological Diagnostic Model For Lateral Lymph Node

Objective:(1)Part one of this study aimed to evaluate the safety and feasibility of laparoscopic lateral lymph node dissection(LLND)for middle-low lying rectal cancer patients and the long survival outcomes.(2)Part two of this study was designed to describethe the variation of visceral branches of internal iliac artery and the relationship between lateral lymph node(LLN)distribution and these vasculars.(3)Part three of this study was designed to explore the safety and feasibility of modified technique of laparoscopic LLND combining fascia-oriented dissection and routine resection of visceral hypogastric vessels for mid to low-lying rectal cancer.(4)Part four of this study aimed to achieve node-by-node matching for LLN on radiologic imaging and pathological examination,and further explore radiological features related to lateral lymph node metastasis(LLNM).(5)Part five of this study aimed to extract and select texual features on CT imaging of targeted LLN to construct diagnostic model for LLNM.Method:(1)A retrospective review was performed in the prospective colorectal cancer database of researcher’s hospital,to find patients who received LLND at the primary surgery from January 2009 to September 2018.The patients were divided into laparoscopy and open group according to the surgical approach.The primary endopoint was postoperative complications according to Clavin-Dinho classification;and the secondary endopoint was overall survival(OS),disease-free survival(DFS),and local recurrence rate.Cox regression was used to explore the risk factors related with survival;and logistical regression was applied to find the risk factors for LLNM.(2)Mid-low lying rectal cancer patients with visible LLN(short diameter≥2mm)on primary CT imaging were enrolled in this prospective study.Then CT angiography would be given to show the variance for the visceral branches of internal iliac artery.The visceral branches included superior/inferior vesical artery,vaginal artery,middle rectal artery,uterine artery,and obturator artery.The lateral pelvic compartment was sub-divided into eight sections,listed as follow: common iliac,external iliac,interal iliac(proximal,distal,and most-distal),obturator(cranial side,caudal side,and most-distal).The size and distributation of LLN and the distance between LLN and adjacent vessels would be calculated and analysed.(3)Mid-low lying rectal cancer patients with swollen LLN(short diameter≥5mm)on primary CT/MRI imaging was enrolled in this prospective cohort study from October 2018 to November 2020.CT angiography and 3D reconstruction for LLN and vasculars would be given before surgery.During operation,the modified LLND procedure was peroformed.The primary endopoint was the rate of postoperative complication(Clavin-Dinho Ⅲ-Ⅳ).The secondary endopoint was the number of LLN harvested and positive rate of LLN.The outcomes would be compared with those of patients receiving the TME alone procedure within the same period and those from previous LLND series.Logistic regression would be applied to find the risk factors for urinary dysfunction after surgery.Kaplan curve would be drawed for survival outcomes and CUSUM method would be used to calculate learning curve for the modified LLND technique.(4)Based on the preoperative CT angiography and 3D reconstruction,we would label the important vasculars and tissues during LLND procedure,according to the relationship between LLN and blood vessels.After surgery,the targeted LLN was carefully picked out on the fresh specimen accoding to the labels,to achieve node-by-node matching.The targeted LLN would be sent separately for pathological evaluation.If the targeted LLN was not found on the fresh specimen,the“node-by-node” matching would be counted as a failure.The pathological and radiological features of the targeted LLN would be collected and analysed.(5)The region of interest(ROI)of the targeted LLN would be drawed on CT imaging,and textual features would extrated(3D slicer,https://www.slicer.org/,version 4.10.2).The logistic univariate analysis and Pearson’s correlation coefficient would be used to select features.Then nomogram model for LLNM would be constructed and validated.Results:(1)From January 2009 to September 2018,a total of 67 patients who received TME and LLND were included in this study,and among them 57 received laparoscopic procedure and 10 received open procedure.Six patients(10.5%)in laparoscopic group developed major complication,including 1 anastomotic leakage,1anastomotic stricture,3 lymphocele,and 1 pulmonary embolism.Minor complication occurred in 19 patients in laparoscopic group,among whom 4 patients needed catheter re-inseration.No perioperative mortality occurred.No signuificant difference was observed between laparoscopic and open groups with respect to the rate of postoperative complication.With the median follow-up time of 45(1-124)months,the DFS for laparoscopic and open groups was 73.7% and 60.0% respectively(P=0.479),and OS was 80.7%and 70.0% respectively(P=0.951).Lateral pelvic recurrence was found in one patient in laparoscopic and two in open group.The DFS(P=0.008)and OS(P=0.025)for patients with LLNM were significantly worse than these without LLNM.Cox multivariate analysis demonstrated that LLNM was one independent risk factor for DFS(HR=3.889;95%CI(1.301～11.689);P=0.015)and OS(HR=6.169;95%CI(1.334～28.539);P=0.020).Besides,logistic multivariate analysis showed N2 stage(mesorectal lymph node metastasis)was the only risk factor for LLNM(OR=40.990;98%CI(2.616～642.359),P=0.008).(2)From October 2018 to November 2020,a total of 155 patients were enrolled,among them 105 were male patients.With the summarizing for the visceral branches of internal iliac artery in 310 lateral sides,we found superior vesical artery and uterine artery would most often originate from umbilical artery at the proximal end.Inferior vesical artery in male had similar origination with vaginal artery in female.Inferior vesical artery in female most often originated from vaginal artery.Besides,middle rectal artery mostly originated from inferior vesical artery in male and vaginal artery in female.A total of 537 LLN(short diameter≥2mm)was detected on the primary CT imaging.The median number of LLN for each patient was 3(1-14).The number of LLN was significantly related to the status of mesorectal fascia(MRF)on primary imaging.A total of 132 LLNs situated in the internal iliac area,including 17 in proximal,92 in distal,and 23 in most distal,among them 62 were with a short diameter ≥5mm.LLNs(66.7%)in internal iliac area mostly distributed around inferior vesical artery and internal pudendal artery.In contrary,few LLNs were detected around umbilical artery/superior vesical artery/uterine artery or superior/inferiror gluteal artery(5/132).278 LLNs were found in boturator area,including 85 in cranial side,169 in caudal side,and 24 in most distal part,and among them only 60 LLNs were larger than 5mm on short axis.For LLNs in the caudal and most-distal end of obturator area,most were also distributed around inferior vesical artery and internal pudendal artery(80/193).The nearest distance from LLNs in most-distal obturator region to the connecting line of internal iliac artery was 2.5(1.2-6.0)mm.According to the vascular variation and LLN distribution we classified internal iliac artery into several categories and listed as following: I the common trunk of internal iliac artery;II superior gluteal artery(the posterior trunk of internal iliac artery);III internal pudendal arterty /inferior gluteal artery/the common trunk of the two arteries(the anterior trunk of internal iliac artery);IV umbilical artery/superior vesical artery/uterine artery;V inferior vesical artery/vaginal artery/middle rectal artery;VI obturator artery.(3)From October 2018 to November 2021,the modified LLND technique was successfully performed in 67 patients.The median time and blood loss for LLND procedure was 80 min and 30 ml respectively.Nine patients developed major complication(Clavin-Dinho Ⅲ-Ⅳ)after surgery,including 6 lymphocele.The rate of other major complications was not significantly different between LLND and TME alone group.Urinary dysfunction occurred in 29 patients in LLND group,and only seven of them needed catheter re-inseration,which was not significantly different from that in TME alone group.The multivariate analysis demonstrated that increased blood loss was the only independent risk factor for urinary dysfunction.The median number of LLN harvested on a single sidewas 10(1-34).Pathological LLNM was detected in 27 patients,and one out of 13 patients receiving bilateral LLND had bilateral LLNM.Within the median follow-up time of 12(3-27)months,12 patients developed recurrence and no recurrence in lateral pelvic compartment was observed.3 patients died of disease progression.The sexual recovery was not significantly different between LLND and TME alone group.The learning curve of the modified LLND technique peaked at the 25 th procedure.(4)Node-by-node matching was successfully achieved for 113 LLNs.Among25 LLNs of interest receiving no neoadjuvant therapy,12 were confirmed positive pathologically.For the other 88 LLNs in patients received neoadjuvant treatment,87 were from patients received neoadjuvant chemoradiotherapy(neoCRT)and one from patients received neoadjuvant chemotherapy(neoCT).26 out of the 88 LLNs after neoadjuvant treatment were pathologically confirmed positive.Among the 62 pathologically negative LLNs after neoadjuvant therapy,12 LLNs were detected with muscin deposit,hence were considered positive before treatment,other 12 LLNs were shown to have fribosis,hence were considered undetermined as presentation,and the other 38 were shown to have normal lymph node features(negative at presentation).Further analysis for the 101 LLNs which the primary status coule be determined demonstrated that AUC value of short diameter for primary LLNM was 0.755(95%CI: 0.655～0.855,P<0.001);the cutoff value was 7mm(sentivity=0.58;specifity=0.90).Besides,diameter on long axis and area were significantly correlated with short diameter.Other parameters,including border,shape,and heterogeneity were not significantly different between primary positive and negative LLNs.The subgroup analysis based on pathological status for the 87 LLNs after neoCRT demonstrated that the AUC value for pre-treatment short diameter,post-treatment short diameter,and shrinking degree was 0.810(95%CI:0.706～0.914,P<0.001),0.887(95%CI: 0.806～0.969,P<0.001),and 0.765(95%CI:0.643～0.888,P<0.001)respectively.Among 19 LLNs of which short diameter<5mm at presentation,none was detected positive pathologically after neoCRT.However,still 4 positive LLNs were detected after neoCRT in all LLNs of which the post-treatment short diamenter was less than 5mm.(5)ROI was drawed for LLNs of interest and textual features were extracted,selected and fitted as Radscore.Radscore was the only significant parameter in logistic multivariate analysis.Thus,nomogram model for LLNM was constructed based on Radscore.For primary LLNM,the AUC value was 0.854(95%CI:0.785～0.927;P<0.001)in the training group and 0.829(95%CI:0.709～0.946;P<0.001)in validatiobn group.For pathological LLNM after long-course neoCRT,the AUC value was 0.942(95%CI:0.888～0.996;P<0.001)in training group and0.872(95%CI:0.716～1.000;P=0.004)in validation group.Conclusion:(1)Laparoscopic LLND is clinical safe and feasible.However,more attention should be paid to the lateral recurrence after LLND.Modified LLND technique and enhanced systematic neoCRT should be dedeveloped in hope for better cancer clearance.(2)The variations in visceral branches of internal iliac artery,with respect to their origin and coused were wide.CT angiography can provide more information on the relationship of LLNs and blood vessels,and is helpful in the LLND procedure.Most LLNs in the distal internal iliac region and the distal obturator region locate around or near to the inferior vesical artery/vaginal artery/middle rectal artery,very fewer LLNs appear around umbilical artery/superior vesical artery/uterine artery,this is important for the procedure of LLND.(3)The modified technique of laparoscopic LLND combining fascia-oriented dissection and routine resection of visceral vessels is safe and feasibile.Compared with conventional LLND,it can achieve better protection for pelvic plexus,increase LLNs harvested,and does not increase the incidence of short-term major postoperative complications.It also does not bring additional damage to long-term urinary function when compared with TME surgery alone.Besides,the modified technique provides better surgical landmarks and plane and reduces the learning curve for beginner.The effect on sexual recovery and survival still needs larger simple size and longer follow-up time.(4)Node-by-node matching is technically feasible for LLNs.It could help further improve the diagnostic accuracy for LLNM.(5)CT-based texual features of LLN are useful to predict LLNM.It has significant advantages when compared with short diameter and some other parameters.