This report describes our single-center experience with a multimodality approach comprising salvage surgery and IORT for the treatment of rRPTs. Notably, our analysis demonstrated satisfactory long-term LC rates, particularly for cases in which an R0 resection was achieved.
Our study included several histological tumor types and thus demonstrated the different clinical scenarios wherein IORT can be employed. In a study involving 128 patients, Cambeiro et al. [
9] also analyzed the results of IORT for the treatment of several types of tumors, such as soft tissue sarcomas as well as head and neck, uterine, and colorectal tumors. The authors used a treatment regimen similar to that used in our study, with association of EBRT (median dose 46 Gy) and IORT (median dose 15 Gy) in 58% of the cases. In a multivariate analysis, only the degree of resection statistically significantly influenced LC, such that the patients who underwent R2 resection had a 2.2-fold higher risk of treatment failure (95% CI: 1.2–4.1;
p = 0.007). In that study, patients with histologically epithelial tumors had lower survival rates. By contrast, in our study, we observed worse survival outcomes associated with sarcoma histology, which we attribute to the presence of this histology in tumors with more aggressive clinical behavior, such as recurrent retroperitoneal and uterine sarcomas. In relation to dose intensity, the authors concluded that the treatments performed with a combination of IOERT and EBRT, with a median value for tumor control standardized to 2 Gy equivalents (EQD2) ≥ 62 Gy had a statistically significant influence on overall survival, compared to treatments with isolated IOERT, with a median value of EQD2 of 31.2 Gy (HR 2.2, 95% CI: 1.1–4.1). In our study, the median value of the EQD2 for the treatments performed with the association of the two modalities and with isolated IORT were 66.2 Gy and 31.2 Gy respectively, but we did not verify influence on overall survival, probably due to the restricted number of patients and by the limitations of a retrospective study.
Below, we further discuss the outcomes of our study in terms of the type of recurrence.
Recurrent colorectal tumors
Recurrent colorectal cancers present unique challenges in terms of management, as previous EBRT limits the available options for further radiation treatment. However, IORT may be indicated in select cases.
Previous studies of IORT have reported variable LC and OS outcomes. One extensive IORT review demonstrated a significant improvement in LC among patients with locally recurrent pelvic tumors [
10]. Furthermore, in a meta-analysis of > 3000 patients, Mirnezami et al. observed benefits with IORT in terms of LC, disease-free survival, and OS among patients with recurrent rectal and advanced colorectal cancers [
11]. A Cleveland Clinic study reported a 1 year LF rate of 16% and 3 year OS rate 49% among patients who underwent IORT for locally advanced and recurrent rectal cancers [
12], and Roeder et al. reported 5 year LF and OS rates of 59 and 30%, respectively, in a similar clinical situation [
3]. Hyngstrom et al. reported 1, 3, and 5 year LF rates of 16, 40, and 44%, respectively, with high-dose-rate intraoperative brachytherapy [
13].
In one case involving a recurrent colorectal tumor in the lateral pelvic wall, we performed intraoperative brachytherapy using a HAM applicator because the anatomical tumor bed location did not allow adequate electron-beam therapy.
Several previous studies have identified surgical resection margin status as an important factor affecting LC during treatment of patients with local recurrences of colorectal cancer. In a recent clinical review, Haddock MG [
14] extensively analyzed several studies and demonstrated that in selected series involving IORT for the treatment of recurrent rectal cancer, the LC rates were 60–80% and the 5-year survival rates were 40–50%. Among cases with microscopically positive margins, the LC rates ranged from 30 to 60% and the 5-year survival rates ranged from 20 to 30%. Other studies have also evaluated the influence of resection grade on the LC of recurrent colorectal tumors. For example, in a study by Roeder et al., patients with a status of R0 after the resection of locally recurrent rectal cancer were found to have 5 year LC and OS rates that were respectively threefold and fivefold better than those of patients with incomplete resections [
3]. Consistent with that report, three resections of recurrent rectal tumors (25%) in our series achieved an R1 status, and none obtained LC. Holman et al. reported a 5 year LF of 45% in a pooled analysis of 565 patients treated with IORT for advanced and recurrent rectal cancers and identified the resection grade as a risk factor for LC [
15]. Dresen et al. [
16] also observed worse LC rates following R1 resection in an analysis of 147 patients with local recurrences of rectal cancer. In his series, the median OS duration was 28 months, whereas the 5-year OS, DFS, metastasis-free survival, and LC rates were 31.5, 34.1, 49.5, and 54.1%, respectively. Resections R0, R1 and R2 was achieved in 84 (57.2%), 34 (23.1%), and 29 (19.7%) patients respectively. For patients with a resection R0, median Os was 59 months and 5-years OS and LC were 48,4% and 68,9% respectively (
p < 0,001). In our study, these treatments yielded 9 R0 (75%) and 3 R1 resections (25%), with a 5 year LC rate of 73%. Consistent with that report results, none of the resections with R1 status obtained local control.
Despite the favorable LC and OS outcomes of IORT, treatment complications are reported with relative frequency in the literature. In our experience, we observed chronic neuralgia in 3 patients and osteomyelitis in 1 patient. Suzuki et al. reported a 5 year LC rate of 60% among patients who received EBRT before or after adjuvant IORT [
17]; however, this outcome was unfortunately accompanied by a grade 3 toxicity incidence exceeding 30%. Similarly, Willet et al. reported a complication rate of 30% after preoperative EBRT associated with IORT for recurrent pelvic tumors [
18]; here, most events involved soft tissue or sacral injury and pelvic neuropathy. Roeder et al. identified wound healing disturbance as the most common complication (20% of patients), followed by abscess or fistula formation (16%) and severe chronic pain (8%) [
3].
Finally, most relevant literature reports have reported DF with this type of neoplasia. Consistent with those reports, half of the 12 cases of recurrent colorectal tumors in our study developed a DF, including 6 cases of liver metastasis.
Recurrent retroperitoneal sarcomas
Approximately 38% of all sarcomas arise in the retroperitoneum [
19], and surgery is currently the main treatment with curative intent. In contrast to sarcomas of the extremities, however, wide surgical margins are often not achievable in the retroperitoneum; accordingly, local progression is the dominant pattern of failure. As the retroperitoneum contains many critical organs with low radiation tolerances, adequate dose delivery via EBRT either cannot be achieved or would result in excessive toxicity. By contrast, IORT facilitates the delivery of a single high radiation dose to the tumor bed during surgery while sparing the surrounding organs at risk via physical distance from the radiation field or adequate lead shielding. The total combined dose from IORT plus moderate EBRT increases the possibility of LC while reducing toxicity.
Regional failure is commonly observed among patients with retroperitoneal sarcoma, including 5 patients in our study; an additional 5 patients experienced DF.
No randomized data regarding radiotherapy for retroperitoneal sarcoma are currently available. An ongoing EORTC study (NCT01344018), Surgery with or without Radiation Therapy in Untreated Nonmetastatic Retroperitoneal Sarcoma (STRASS), is currently recruiting participants. Furthermore, propensity score-matched analyses of more than 9000 patients with resected retroperitoneal sarcoma with or without preoperative or postoperative irradiation in the National Cancer Data Base demonstrated improved median OS in the irradiation group compared with the no-irradiation group, regardless of the irradiation time (110 months for preoperative vs. 89 months for postoperative vs. 66 months for no irradiation) [
20].
Regarding LC, other authors have reported results similar to ours. Petersen et al. evaluated the management of 44 cases of recurrent retroperitoneal soft tissue sarcoma treated with IORT at the Mayo Clinic; the reported LF rate of 39% was comparable with our series. The authors also reported a 5 year OS rate of 48%, with no differences were found by primary or recurrent tumor status [
21].
Hager et al. compared two groups of patients; although all underwent surgery, half also received radiation therapy [
22]. As in our study, Hager and colleagues typically administered a median IORT dose of 15 Gy using an electron energy of 6 MeV. The combination of surgery and IORT significantly improved survival outcomes, compared with surgery alone (
p = 0.04). For all patients, resections R1 and R2 resection have a decreasing in 5 years survival rate by 7.6% and by 34.7% respectively compared to R0 resection.
Recently, Roeder et al. [
23] published a retrospective study involving 156 patients among whom 87 had recurrent tumors. Total 114 patients were treated with a combination of IORT (median dose: 15 Gy) and EBRT (median dose: 45 Gy). During a median follow-up of 38 months, the LC rates at 3 and 5 years were 57 and 50%, respectively. In a multivariate analysis, the tumor grade resection margins and the association with EBRT remained statistically significant.
Recurrent gynecological tumors
Patients with recurrent gynecological tumors (e.g., tumors of the uterine cervix, endometrium, and ovary) often present with lesions on the pelvic walls and/or involvement of the pelvic or paraaortic lymph nodes. Again, a previous treatment history of high-dose EBRT for these tumors limits the options for salvage radiation therapy. Accordingly, IORT is an important therapeutic option.
IORT has been studied for the treatment of recurrent gynecologic tumors since the 1990s. In a Mayo Clinic study of 148 patients, 125 patients had recurrent gynecological tumors and 113 received IORT associated with EBRT [
24]. In that study, the 5 year LF rate was 40% and the 5 years OS was 27%. Furthermore, R2 resection was associated with a worse 5 year OS rate, compared with R0 or R1 resection (31% vs 13%,
p = 0.01).
Tumors of the uterine cervix are common in Brazil and are often diagnosed at advanced stages. Some retrospective studies in the literature have evaluated the use of IORT for the treatment of recurrent uterine cervical cancer. Tran et al. evaluated 17 patients treated with orthovoltage IORT at a median dose of 11.5 Gy [
25]. In that study, the LC, metastasis-free survival, and specific survival rates were 45, 60, and 46%, respectively. Similar to our treatment scheme, a Spanish study reported the outcomes of 36 patients with recurrent primary cervical tumors who were treated with IORT (median dose: 15 Gy) alone or combined with EBRT (45 Gy at 1.8 Gy per fraction) if prior radiotherapy had not been administered [
26]. In that study, the 10 year LC was 47%, and factors that adversely affected LC included parametrial margin involvement, R1 resection, and pelvic lymph node involvement.
The Mayo Clinic reported the outcomes of 25 patients treated with IORT for recurrent endometrial cancer [
27]. Most patients presented with involvement of the pelvic sidewall or paraaortic nodes. In that study, 21 of 25 patients received EBRT (median dose: 45 Gy). The median IORT dose of 15 Gy was consistent with our study. The median survival duration was 57 months, and the 5 year OS rate was 47%. The authors of that study reported that the resection grade influenced survival, with 5 year OS rates of 71, 40, and 0% among R0, R1, and R2 cases, respectively. LF within the IORT field and DF were observed in 4 (16%) and 6 patients (24%), respectively. Given the small number of patients with recurrent gynecological tumors in our study, we could not analyze the influence of histological type on survival. However, other studies have reported better survival outcomes for recurrent endometrial tumors treated with IORT. In a study involving 36 patients with recurrent gynecological malignancies conducted by Arias et al. [
28], an endometrial histology was found to correlate with better rates of local PFS (
p = 0.017) and OS (
p = 0.038). Furthermore, older patients exhibited significantly better distant PFS outcomes (
p = 0.015), and patients with endometrial cancer tended to better distant PFS relative to patients with recurrent cervical and vulvar tumors.
Investigators at Stanford University reported the use of orthovoltage IORT to treat 22 patients with recurrent ovarian cancer [
29]. In that study, a median dose of 12 Gy was administered to various sites, including the pelvis, paraaortic nodes, inguinal nodes, and porta hepatis. Nine, 5, and 6 patients received whole-abdominal EBRT, loco-regional EBRT, and associated chemotherapy, respectively. The median survival duration was 26 months, and the 5 year OS and disease-free survival rates were 22 and 18%, respectively. LC and RC was 68% at 22 months. Furthermore, 55% of patients experienced DF; consistent with our study in which 9 patients (60%) experienced DF, the authors identified this failure pattern as an important influence on OS.
For the various types of tumors, we observed a direct influence of the surgical margin status on LC. Some authors believe that R1 or R2 resection may be a consequence of selection of tumor variants with more aggressive biological behaviors following the initial course of radiotherapy and chemotherapy [
16]. Despite progress in the quality and precision of imaging exams and the rigorous selection of clinical cases, a true evaluation of the possibility of obtaining a R0 surgical margin may only be possible during surgery. In our experience, we considered cases with increased risk and those with tumors very close to bone structures, such as recurrences of pelvic tumors in the pre-sacral space, recurrences of retroperitoneal sarcomas in the paravertebral region, and lesions adhered to vascular structures. Such cases may be considered “borderline” and should be the subject of a more extensive evaluation to determine the indications and contraindications for IORT. The intraoperative identification of R1 or R2 margins might indicate the need for a dose increase, which is often difficult to accommodate depending on the previously administered EBRT dose and the presence of high-risk anatomical structures in the tumor bed. Other alternative treatment techniques for recurrences of pelvic tumors have been reported in the literature. Murray et al. [
30] analyzed 17 studies of stereotactic ablative radiotherapy (SART) for the treatment of 205 patients with recurrent malignant disease within the pelvis. SART has the advantage of allowing the previous treatment planning, optimizing the coverage of the target volume of treatment. The treatment being performed in a fractionated dose regimen allows the radiobiological advantage of better protection of normal structures.
The authors reported 1-year CL indices ranging from 51 to 100% and a low rate of grade 3 or 4 complications.
Morbidity
Surgeries performed for the salvage treatment of pelvic or retroperitoneal tumor recurrences are extensive, debilitating, and potentially complicated by previous treatments. First-line EBRT induces fibrosis in tissues surrounding the recurrent tumor and increases the difficulty of the surgical procedure. We observed some serious complications; however, they were fortunately not fatal. Of these complications, 70% occurred within 90 days of surgery and could be attributed to the surgical procedure.
Peripheral neuropathy is the most commonly reported toxicity attributed to pelvic IORT. IORT-related neuropathies usually present with pain but no significant motor or sensory loss. The pain is usually chronic, possibly severe, and is often manageable with analgesics, including opioids. In our study, this complication occurred in 7 patients and mainly affected those who received treatment for recurrent pelvic region tumors (5 cases). In such cases, pain was likely attributable to the presence of nerve structures in the treatment field (i.e., very near the tumor bed); here, nerves cannot be shielded and thus receive the full IORT dose.
Some authors have attempted to correlate the incidence of neuropathy with the IORT dose. A Mayo Clinic analysis involving 51 patients treated with IORT (doses: 10–25 Gy) in association with EBRT for primary or recurrent pelvic tumors observed grade 1–3 neuropathy in 32% of patients [
31]. Haddock et al. analyzed the relationship between the IORT dose and incidence of neuropathy in patients receiving treatment for recurrent rectal tumors and noted that IORT doses of ≥12.5 Gy were associated with an incidence of 5% grade 2 or 3 neuropathy, whereas doses > 15 Gy were associated with 14% incidence of grade 2 or 3 neuropathy [
32]. Of the 7 patients who developed pain in our study, 6 received an IORT dose ≥15 Gy (range: 15–20 Gy). However, given the small number of patients with this complication, it was not possible to establish a statistical association between the IORT dose and the incidence of neuropathy.