Although the safety of laparoscopic surgery for colon cancer had been demonstrated in several studies [4-7], the safety of this surgical approach is controversial in T4 colon cancer. Several studies have suggested that a laparoscopic approach in T4 colon cancer may be feasible in some patients. Few studies have provided useful indications for laparoscopic surgery in T4 colon cancer. Klaver et al.  reported that laparoscopic surgery for T4a tumors might be safe. However, the pathologic features would not be helpful in determining the indication of laparoscopic surgery preoperatively. Park et al.  found the laparoscopic approach to be feasible for left-sided T4 colon cancer. Nevertheless, a useful predictor is still necessary to preoperatively determine the safety of laparoscopic surgery for T4 cancer.
In this study, the clinicopathologic and oncologic outcomes of laparoscopic surgery for T4 colon cancer were generally comparable to those of open surgery. The laparoscopic approach, especially for small T4 tumors, had better 3-year DFS rates than open surgery. To adjust for confounding variables, we analyzed the Cox proportional hazards regression model for OS and DFS in the entire cohort and in patients with tumor size ≤4.0 cm. Laparoscopic surgery had better DFS rates in patients with tumor size ≤4.0 cm (p = 0.020) (Additional Files 1 and 2).
A previous study  has reported that malignant cells are intraoperatively exfoliated from the tumor during resection and spread to the peritoneal surface and portal vein system. This can be prevented by minimizing tumor manipulation, e.g., through laparoscopic surgery. Lacy et al.  showed better cancer-related survival with laparoscopic colectomy than open surgery for non-metastatic colon cancer in a randomized clinical trial, as did our study. When laparoscopic surgery is conducted by an experienced surgeon, tumor spillage and spread may be prevented in some patients.
As tumor size increases, some technical challenges arise with regard to laparoscopic surgery, because it reduces the working space, narrows the operative visual field, increases bleeding, and makes the tumor difficult to remove. Moreover, larger tumors increase the risk of tumor spillage, thereby increasing peritoneal seeding or trocar-site recurrence. Our data show that the 3-year OS and DFS rates in patients with tumor size >4.0 cm are not significantly different between the two groups (84.4 vs. 87.8%, p = 0.22 and 80.6 vs. 68.7%, p = 0.091, respectively), suggesting that the laparoscopic approach is more feasible in patients with small tumors than in those with larger tumors.
Laparoscopic surgery is better than open surgery in regard to perioperative outcomes. In previous studies [1, 16, 17] comparing laparoscopic and open surgery in T4 colon cancer, laparoscopic surgery was associated with less intraoperative blood loss, which has been proven to be a predictor of long-term survival [18, 19]. Some studies [20, 21] have shown that hospital stays are shorter in patients who undergo laparoscopic surgery. In this study, patients in the laparoscopic group had less intraoperative blood loss and shorter hospital stays than those in the open group.
In a previous study  of T4 colon cancer, the conversion rate from laparoscopic to open surgery was reported to be in the range of 7.1–28.2%. Converted patients have high postoperative morbidity and adverse effects on long-term oncologic outcomes . In the present study, the overall conversion rate was 7%, and the conversion rate for patients with tumor size ≤4.0 cm was 2.3%. The low conversion rate might be responsible for the better oncologic outcomes of laparoscopic surgery.
In this study, the 3-year DFS rate of patients in the open group with tumor size ≤4.0 cm was 53.2%, which was much lower than the 75.1% for all patients in the open group. This result is similar to that of the study by Huang et al. , which reported that a smaller tumor size was associated with a decreased survival in the T4b subset of colon cancer patients. Huang et al.  suggested that small tumors in T4b patients may reflect a more biologically aggressive phenotype. Another plausible explanation is that surgeons may have conducted more aggressive surgery for larger tumors. In the present study, the rate of multi-visceral resection was 28.2% in the entire open group, but only 6.9% in the small tumor group. Although R0 resection was accomplished in all patients with small tumors, it is possible that disseminated lesions remained in adjacent organs. These may have contributed to the worse 3-year DFS rate in patients with tumor size ≤4.0 cm in the open group.
The limitations of this study are as follows. As this was a retrospective study, the choice of surgical approach may have been influenced by the patient’s condition or tumor progression. First, this study was conducted on the basis of the pathological T4 instead of the clinical T4, although the former cannot be used to determine the surgical approach preoperatively. Engelmann et al.  reported that the computed tomography accuracy of T4 staging in colon cancer was only 70–77%, although further studies are needed in patients with clinical T4 colon cancer. Second, more patients had higher ASA scores in the open group. This may have affected OS or DFS. However, in patients with tumor size ≤4.0 cm, there was no intergroup difference in ASA scores. Third, the T4b rate and number of adjacent organ resections were higher in the open group. Thus, it is apparent that open surgery was chosen for patients with more advanced tumors. However, there were no intergroup differences in the T4b rate and number of adjacent organ resections in patients with tumor size ≤4.0 cm.