This study was performed prospectively in the First Affiliated Hospital of Naval Military Medical University June 2018 to December 2018. The experiment was approved by Ethics Committee of the First Affiliated Hospital of Naval Military Medical University and in accordance with precepts established by the Helsinki Declaration. All patients signed informed consent before surgery. We included the total cohort of 20 consecutive patients who fulfilled the following inclusion criteria: (1) older than 18 years; (2) diagnosed with rectal cancer; (3) had already undergone preoperative radiotherapy; (4) had an intention to receive radical resection. The exclusion criteria: (1) refused to receive operation after radiotherapy; (2) clinical complete response and received “wait and watch” approach; (3) local resection after radiotherapy; (4) patients who did not finished the radiotherapy. All patients received intensity-modulated radiation therapy in the case of preoperative stage T3-T4 or N+, or if the circumferential margin was considered positive: a total dosage was 45 ~ 50.4Gy (1.8 ~ 2.0 Gy per time, 25-28 fractions). The preoperative treatment was performed at the Department of Medical Oncology. Afterwards, the radical resection including low anterior resection (LAR) or abdominal perineal resection (APR) was given to the patients at the Department Colorectal Surgery based on the evaluation of response of the tumors to the preoperative radiotherapy.
Determination of the tumor visible margin in vivo with naked eyes
The size of the tumor, the distal tumor visible margin was determined intra-operatively based on morphological abnormalities by palpation and surgeons’ naked eyes. This morphological abnormalities can be presented as mucosal fold changes, poor mobility and hardening texture as a fibrotic scar tissue, characterized by red color, swelling or ulceration.
Labeling of the tumor visible margin with sutures and nanocarbon particles
Directly after resection, after reaching agreement among the same group of colorectal surgeons about the position of distal visible margin, the colorectal surgeons labeled the distal visible tumor margins at distal 6, 5, 7 o’clock positions using a 5-0 suture. After being labeled with suture, the specimen was fixated in 10% formalin for 24 hours (Fig 1a, b). The fixed surgical specimens were injected with fewer than 0.02 ml carbon nanoparticles suspension (Lummy, Chongqing, China), at the 5, 6, 7 o’clock positions, as indicated by the marking sutures. An gauge 27 skin test syringe was utilized to decrease the influence on the results from diameter of the needle. We replaced the sutures with injecting carbon nanoparticles because sutures would be lost during cutting. The carbon nanoparticles was injected perpendicular to the surface of the specimen, through the layer of mucosa, submucosa, and muscle layer to mark the distal visible margin of the tumor according to the naked eyes for facilitating identification of tumor intramural spread distance under the microscopy. Both the resection margin connected to small amount of tumor tissue as comparison were collected in 2 cm-long blocks. From each rectal specimen, three tissues samples were collected at 5, 6, 7 o’clock position along the nanocarbon markers (Fig 1c). These three samples from each patient were used to investigate the intramural tumor spread distance under the microscopy in the following procedure.
Microscopic evaluation of the tissues
All the samples were embedded in paraffin using Paraffin Embedding Module (Leica EG1150, Leica, Germany) and cut into sections (4 µm thickness) in the longitudinal direction. The microscopic evaluations were performed using a light microscope (Leica DM4 B, Leica, Germany) connected to a digital camera (Leica DFC9000, Leica, Germany). The microscopic evaluation was performed only ex-vivo, i.e. on the resected tissues. The hematoxylin and eosin staining was carried out to evaluate the morphology of the tissue samples.
Estimation of the distal intramural spread distance
The steps made to measure the distal spread distance are shown in Fig 2. The intramural spread distance was defined as the distance between visible margin and microscopic tumor margin. Ex vivo intramural spread distance i.e. distance between the carbon nanoparticles and distal tumor cells was measured ex vivo under the microscopy for each slide. Firstly, the carbon nanoparticles indicating the visible tumor margin under naked eyes was used as a zero point (Fig 2a, black line as visible margin, step 1). Secondly, based on the morphological view, the tumor cells located distally were found, and this way the distal tumor margin under microscopy (microscopic margin) was determined (Figure 2a, red line as microscopic margin, step 2). Lastly, the distance between visible margin (black line) and microscopic margin (red line) was measured as ex vivo distal intramural spread distances (Figure 2a, red double arrow length as ex vivo spread distance, step 3). In case the tumor cells were located proximally to the carbon nanoparticles, the distal spread distance was described as negative (Figure 2b, d). If no tumor cells were found and only mucus was found, the value was 0 cm. The distal intramural spread distance in vivo was the value ex vivo multiplied by 1.75 as shrinkage factor according to literature (11). The tumor differentiation level, tumor deposits, tumor budding, perineural invasion, tumor regression grade and T/N-stage were also recorded. Data acquisition was performed with ImageScope v12.2.2.5015 software (Aperio, USA).
Statistical analysis
All statistical analysis was performed using IBM SPSS Statistics, Version 23.0 (Armonk, NY, USA: IBM Corp.). Continuous values were reported as mean +/- standard deviation (SD) or median +/- interquartile range (IQR), depending on whether the data were normally distributed or not. Categorical values were reported as absolute numbers and percentages. Continuous variables were compared using t-test for independent samples, if the variables were normally distributed or Mann-Whitney test, if not normally distributed. Categorical variables were analyzed by using the chi-square test. A P-value <0.05 (two sided) was considered statistically significant.