Neoadjuvant chemotherapy with modified FOLFOXIRI for locally advanced rectal cancer to transform effectively EMVI and MRF from positive to negative: results of a long-term single center phase 2 clinical trial

Over 1.9 million new colorectal cancer cases and 915,880 deaths were estimated to occur worldwide in 2020 [1]. Overall, colorectal cancer ranked third in terms of its incidence and second in terms of mortality [1]. New cases of rectal cancer accounted for 38.9% of all colorectal cancers [1]. In China, the incidence of colorectal cancer was ranked third in 2015 and its incidence is currently increasing [2]. The proportion of rectal cancer patients is significantly higher in China than in Europe and the US, although had declined from 72.6% in the 1980s to 49.6% in 2013 [3, 4].

Fluorouracil-based preoperative chemoradiotherapy (CRT) followed by total mesorectal excision (TME) and adjuvant chemotherapy remains the standard treatment for locally advanced rectal cancer (LARC) [5]. In spite of improvements in local disease control, no decrease in the incidence of distant metastases was observed and the beneficial effects on overall survival (OS) remains unclear [6, 7]. The results of phase 3 clinical trials of total neoadjuvant therapy (TNT) reported more significant tumor regression and a reduced incidence of distant metastases, hence TNT has become another standard treatment option. All these studies had high completion rates of chemotherapy over 95% in common, even when the FOLFIRINOX triple regimen was adopted in the PRODIGE 23 study [8]. Furthermore, TNT provided a significantly greater pathological complete response (pCR) rate compared to CRT, which were 27.8% vs 12.1% in the PRODIGE 23 study [8], and 28.4% vs 14.3% in the RAPIDO studies, respectively [9]. In addition, a watch and wait strategy may be implemented if clinical complete regression (cCR) was achieved, which makes changes in the surgical procedure and organ preservation possible.

After consideration of the long-term side effects elicited by radiotherapy, several studies have endeavored to explore the role of neoadjuvant chemotherapy (nCT) as monotherapy. In the FOWARC study, though the neoadjuvant with mFOLFOX6 alone achieved a postoperative pCR rate of 6.6%, the rates of R0 resection (89.4% vs 90.7%), anal preservation (89.5% vs 84.4%), and 3-year local recurrence (8.3% vs 8.0%) were similar compared to CRT, as well as producing less postoperative complications [10]. The 2021 ESMO congress CONVERT trial enrolled patients without mesorectal fascia (MRF) invasion based on pelvic magnetic resonance images (MRI) at baseline [11]. Four cycles of XELOX regimen chemotherapy showed comparable pCR (11% vs 13.8%) and downstaging rates (40.8% vs 45.6%), with capecitabine-based CRT in a neoadjuvant setting. While the occurrences of preventive diverting ileostomy and peri-operative distant metastasis were significantly reduced in nCT compared to CRT patients (52.3% vs 63.6%, P = 0.008; 0.7% vs 3.1%, P = 0.034, respectively). It seemed nCT alone was promising according to the results reported (vide supra).

Given the low pCR rate with mFOLFOX6 in the FOWARC study, we conducted a phase 2 clinical trial to evaluate the efficacy and safety of mFOLFOXIRI, a more intensive and more active triplet regimen proven to be effective for the treatment of metastatic colorectal cancer [12, 13]. We also further explored the role of nCT alone as neoadjuvant treatment for patients with LARC from December 2015. Patient enrollment ended in March 2019. Herein, the long-term follow-up results are reported.

In the present study, nCT with mFOLFOXIRI therapy for patients with LARC achieved a pCR rate of 4.3%, which did not meet the primary endpoint. The pCR rate was similar to the 6.6% for mFOLFOX6 in the FOWARC study and was lower than the 17.4% of mFOLFOXIRI in the FORTUNE study [15], which was also carried out by the FOWARC team. The low pCR rate in our study may be partly related to the higher proportion of patients with high-risk factors and a heavy tumor burden. Indeed, the baseline characteristics in our study showed 96% node-positive and 50% cN2. Retrospective evaluation found that the positive rates of EMVI and MRF were as high as 78.4% and 56.5%, respectively. However, the baseline data from the FOWARC study included 72.1% node-positive, 26% cN2 and 31.4% MRF invasion patients, all of which were lower than in our study. The baseline characteristics of the FORTUNE study showed 84% node-positive, 45.7% cN2 and 29.2% MRF invasion, while EMVI data were not reported. In addition, there has been no commonly recognized optimal dosage of FOLFOXIRI to date. A phase 1 dose-escalation study in Japan reported maximum tolerated dosages as 85 mg/m² for oxaliplatin, 150 mg/m² for irinotecan and 2,400 mg/m² for 5-fluorouracil [16]. We therefore adopted the same dosage regimens in our study. However, the dosage of irinotecan and 5-fluorouracil were both lower than administered in the FORTUNE study (i.e., 150–165 mg/m² irinotecan and 2,800 mg/m² 5-fluorouracil, respectively), which may be one reason for the lower pCR rate observed in our study.

Neoadjuvant radiotherapy is an effective way to improve local control of LARC. In randomized studies of neoadjuvant CRT versus preoperative radiotherapy alone (FFCD9203 [17] and EORTC22921 [18] studies), the pCR rate of neoadjuvant CRT reached 11.4–14% and the 5-year local recurrence rate was 7.6–8.1%, both of which were significantly superior to preoperative radiotherapy alone. Furthermore, in studies related to TNT strategies (PRODIGE 23 and RAPIDO studies), the pCR rate reached 27.8–28%, and the local recurrence rate was 4.8–8.7%. Though the pCR rate was not as high as expected, our results indicated that nCT with 3 to 5 cycles of mFOLFOXIRI resulted in significant downstaging. Among the 44 patients with node-positive tumor at baseline, 20 were node-negative after completion of nCT based on MRI evaluations. In total, 52.1% of the patients were evaluated as node-positive by MRI after nCT, findings consistent with the pathological negative rate of 50% after surgery. Notably, our results also indicated that nCT with mFOLFOXIRI appeared to be quite effective in eradicating EMVI and transforming patients to a negative MRF. MRF invasion and EMVI-positive are among the top high-risk factors for local recurrence and metastasis in rectal cancer [19, 20]. In the present study, the incidence of MRF invasion revealed by MRI showed a sharp decrease from 56.5% at baseline to 8.7% after nCT, and was confirmed to be 2.2% by pathology evaluations. The incidence of positive EMVI also decreased from 78.4% at baseline to 32.6% after nCT and was confirmed to be 34.8% after pathological assessments. It was further found that patients with lymph node and EMVI status transforming from positive at baseline to negative after nCT achieved significantly longer DFS and OS times compared to those with positive status from the outset of the study. MRF status transforming from positive at baseline to negative after nCT had a trend to prolong DFS and OS times. Moreover, nCT with a triplet regimen transformed high-risk patients into low-risk patients, and achieved comparable survival times to that of low-risk patients at baseline. The importance of identifying persistent EMVI at restaging following neoadjuvant CRT or nCT is not entirely understood, however, it has been reported that patients with EMVI positive tumors at baseline that regress to become EMVI negative at restaging have similar survival outcomes and recurrence rates as patients who were EMVI negative at baseline, and that patients who remained EMVI positive at restaging have shorter disease-free survival [21].

On the other hand, re-evaluation of high-risk factors by MRI after nCT may distinguish patients with different prognosis outcomes. GRECCAR-4 was a randomized phase 2 study in which an nCT plan was established based on the degree of tumor shrinkage after intensive induction chemotherapy (FOLFIRINOX) [22]. In the study, 206 patients with LARC received 4 cycles of FOLFIRINOX. The 5-year DFS and OS rate were significantly improved in good responders (n = 30) compared to poor responders (n = 196). In addition, good responders treated with standard CRT plus surgery had a higher 5-year DFS rate compared to those who received immediate TME (89.5% vs 80%), but the 5-year OS rates were comparable (93.3% vs 90%). Although GRECCAR-4 study had the limitation of a small sample size, especially arm A (n = 16) with good responders without radiotherapy, it revealed that patients with tumor remission and clearance of high-risk factors (such as EMVI) would likely obtain a better prognosis. For such patients, preoperative chemotherapy alone might achieve the same OS as preoperative CRT.

A high completion rate of nCT with mFOLFOXIRI was observed in this study: 96% for 3 cycles and 88% for 5 cycles, with manageable toxicities. The incidence rates of the most concerning grade 3/4 toxicities namely neutropenia and diarrhea were 50% and 12%, respectively. There were no treatment-related deaths. The nCT with mFOLFOXIRI did not bring any additional safety problems to surgery. The incidence rates of anastomotic fistula and postoperative infection were 2.2% and 4.3%, respectively. The rates in this study were similar to those reported in the mFOLFOX alone treatment group in the FOWARC study (7.9% and 7.2%, respectively), but were significantly lower than in the two concurrent chemoradiation groups (19.8% and 16.3%, and 18.1% and 14.8%, respectively).

Nevertheless, 11/46 (23.9%) patients retained ypN2 disease after surgery. Finally, 14 (30.4%) of 46 patients received postoperative radiation with concurrent capecitabine therapy mainly because they had pathologically node-positive (n = 12) or CRM-positive (n = 1) tumors. With a median follow-up time of 51.2 months, the incidence of local recurrences (6.5%) and distant metastases (19.6%) appeared to be relatively low, which was promising taking into consideration the fact that most of the patients enrolled in this study were at high-risk. The incidence of local recurrences in this study was numerically comparable to those of 5-fluorouracil- or mFOLFOX6-based CRT in the FOAWRC study and the TNT group in the RAPIDO study. The three-year DFS and OS rates were 75.8% and 86.8%, respectively. Likewise, the survival outcomes also appeared promising numerically compared with the results of the FOWARC and RAPIDO studies.

The present research had several limitations. First, it is a phase 2 single-arm study with a limited sample size. Second, a higher proportion of patients with MRF invasion and EMVI-positive were enrolled, which did not reflect real-world incidence. Third, EMVI and MRF status were post analysis rather than prospective research purposes.

In conclusion, nCT with mFOLFOXIRI had a significant down-staging effect on LARC. It was effective in eliminating EMVI, and transforming positive MRF to a negative status, with a favorable safety profile. The long follow-up showed promising DFS and OS times, indicating that initial neoadjuvant treatment with mFOLFOXIRI could allow a certain subset of patients to be free from radiotherapy. This was especially true for those patients bearing high risks who be eliminated after nCT in the re-evaluation and who probably obtaining better survival, and thus, warrant further investigation.