Interventional procedure overview of intraoperative electron beam radiotherapy for locally advanced and locally recurrent colorectal cancer
Closed for comments This consultation ended on at Request commenting lead permission
Evidence summary
Population and studies description
This overview is based on 3,144 patients from 2 systematic reviews and meta-analyses, 4 case series and 1 cohort study. However, both systematic reviews included studies that used HDR-brachytherapy. Additionally, there may have been significant overlap in patient populations of the case series because of studies being done in similar institutions over similar time periods.
This is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 7 studies as the key evidence in table 2 and table 3, and lists 48 other relevant studies in table 5.
The 2 systematic reviews and meta-analyses of comparative studies considered 7 studies and 15 studies, respectively. There were 2 RCTs (a third RCT was incorrectly reported in 1 of the systematic reviews) reported in the systematic reviews. All other studies were observational and typically retrospective.
Recruitment periods of the key evidence studies began in 1978 (1 study included in Liu, 2021). Follow-up periods, when reported, ranged from 25.5 to 80 months.
The studies included in the systematic reviews were done in various locations, including several countries in Europe, the US, China, and Japan. Of the other key evidence studies, 1 was done in the US, 2 were done in the US and The Netherlands, 1 was done in The Netherlands, and 1 was done in Germany, Italy, Spain, and The Netherlands.
Most studies included people with rectal cancer and only Haddock (2011) included patients with colon and rectal cancer. Most patients in the Fahy (2021) systematic review had locally advanced rectal cancer; a breakdown between LARC/LRRC was not reported in the Liu (2021) systematic review. Of the other studies, 2 included people with locally recurrent colorectal or rectal cancer, and 2 included people with locally advanced rectal cancer. One study specified that patients were only included in the analysis if they had an R1 resection (Voogt, 2021). Table 2 presents study details.
Study no. | First author, date country | Studies/Patients (male:female) | Age | Study design | Inclusion criteria | Intervention | Follow up |
|---|---|---|---|---|---|---|---|
1 | Fahy, 2021 Various | 7 studies (6 IOERT; 1 HDR-brachytherapy) including 2 RCTs n=422 IORT n=411 Surgery/EBRT | Median ages ranged from 58 to 64 | Systematic review and meta-analysis | Confirmed LARC/LRRC; people who had IORT; comparison in terms of local control between a group receiving IORT and a surgery/EBRT only group; outcomes of locoregional recurrence rate or morbidity, more than 10 patients; clear research methodology; published on or after 1 January 2000. | 6 studies used IOERT, 1 study used HDR-brachytherapy; comparator was surgery/EBRT Dose of IORT ranged from 10 to 25 Gy | NR |
2 | Liu, 2021 Various | 15 studies (12 IOERT, 3 HDR-brachytherapy) including 3 RCTs n=687 IORT n=773 surgery/EBRT | Mean ages ranged from 58 to 66 | Systematic review and meta-analysis | Studies that compared IORT and surgery/EBRT treatment for people with rectal cancer; at least 20 patients. | 12 studies IOERT, 3 studies HDR-brachytherapy; comparator was surgery/EBRT Dose of IORT ranged from 10 to 25 Gy | Study means ranged from 25.5 to 80 months |
3 | Voogt, 2021 The Netherlands | n=373 (total LARC and LRRC) n=263 IOERT n=110 HDR-brachytherapy 255 male;118 female | 278 were younger than 70; 95 were older than 70 | Retrospective cohort study; prospectively collected data | Consecutive patients with LARC or LRRC who had an R1 resection after undergoing intentionally curative surgery in which IORT was delivered by either IOERT or HDR-brachytherapy. | IOERT Comparator was HDR-brachytherapy Dose was typically 10 to 12.5 Gy | NR |
4 | Haddock, 2011 US | n=607 369 male:238 female | Median 62 | Retrospective case series; prospectively collected data | Patients with recurrent colorectal cancer who had IOERT and surgical resection. | IOERT Median dose was 15 Gy | Median for survivors 44 months |
5 | Holman, 2017 US/The Netherlands | n=565 346 male:219 female | Mean 61.5 | Retrospective case series; prospectively collected data | Recurrent rectal cancer and without preoperative distant metastases, including only patients in whom the surgical intent was a gross total resection. | IOERT Dose was typically 10 to 12.5 Gy for R0 or R1 resections and 15 to 20 Gy after R2 resection | Mean for survivors 40 months |
6 | Holman 2016 US/The Netherlands | n=417 248 male;169 female | Mean 59.2 | Retrospective case series; prospectively collected data | Primary locally advanced (T4b) rectal cancer without preoperative distant metastases, locally unresectable for cure at initial presentation. | IOERT Dose was typically 10 to 12.5 Gy for R0 or R1 resections and 15 or higher Gy after R2 resection | Mean 52 months |
7 | Kusters, 2010 Germany, Italy, Spain, The Netherlands | n=605 389 male;216 female | Mean 62 | Retrospective case series; prospectively collected data | Locally advanced rectal cancer without preoperative distant metastases. | IOERT Dose was typically 10 to 12.5 Gy | Mean 62 months |
First author, date, study design | Efficacy outcomes | Safety outcomes |
Fahy, 2021 Systematic review and meta-analysis | Locoregional recurrence Meta-analysis of 7 studies: no statistically significant difference between IORT and surgery/EBRT, OR 0.55, 95% CI 0.27 to 1.14; p=0.11.
| Wound infections Meta-analysis of 4 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.13, 95% CI 0.50 to 2.54, p=0.76 Pelvic abscess Meta-analysis of 3 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.01, 95% CI 0.54 to 1.87, p=0.99 Anastomotic leak Meta-analysis of 5 studies: no significant difference between IORT and surgery/EBRT, OR 1.06, 95% CI 0.51 to 2.18, p=0.88 Reintervention rate Meta-analysis of 3 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.13, 95% CI 0.43 to 2.98, p=0.80 |
Liu, 2021 Systematic review and meta-analysis | 5-year OS Meta-analysis of 9 studies: no statistically significant difference between IORT and surgery/EBRT, HR 0.80, 95% CI 0.60 to 1.06; p=0.189 5-year disease-free survival Meta-analysis of 6 studies: no statistically significant difference between IORT and surgery/EBRT, HR 0.94, 95% CI 0.73 to 1.22; p=0.650 5-year local control Meta-analysis of 14 studies: statistically significantly better with IORT than surgery/EBRT, OR 3.07, 95% CI 1.66 to 5.66; p=0.000
| Abscess Meta-analysis of 6 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.10, 95% CI 0.67 to 1.80, p=0.833 Fistulae Meta-analysis of 3 studies: no statistically significant difference between IORT and surgery/EBRT, OR 0.79, 95% CI 0.33 to 1.89; p=0.600 Wound complications Meta-analysis of 8 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.02, 95% CI 0.52 to 2.02, p=0.948 Anastomotic leakage Meta-analysis of 7 studies: no statistically significant difference between IORT and surgery/EBRT, OR 1.09, 95% CI 0.59 to 2.02; p=0.775 Neurogenic bladder dysfunction Meta-analysis of 3 studies: no statistically significant difference between IORT and surgery/EBRT, OR 0.69, 95% CI 0.31 to 1.55; p=0.369 |
Voogt, 2021 Cohort study | LARC (IOERT n=151; HDR-brachytherapy n=64) OS There was no statistically significant difference in OS between treatments (p=0.989)
Significant prognostic factors were age, time between radiation therapy and surgery, pathologic tumour and lymph node stage, and resection margin. Local recurrence-free survival There was no statistically significant difference in local recurrence-free survival between treatments (p=0.103).
However, in multivariate analysis, treatment modality was borderline statistically significantly associated with local recurrence-free survival: HR 0.504, 95% CI 0.254 to 0.999, p=0.050. Other significant prognostic factors included time between radiation therapy and surgery, pathological tumour stage, and resection margin. LRRC (IOERT n=112; HDR-brachytherapy n=46) OS There was no statistically significant difference in OS between treatments (p=0.747)
Significant prognostic factors were age and pathologic lymph node stage. Local recurrence-free survival There was no statistically significant difference in local recurrence-free survival between treatments (p=0.139).
However, in multivariate analysis, treatment modality was statistically significantly associated with local recurrence-free survival: HR 0.567, 95% CI 0.349 to 0.920, p=0.021. Other significant prognostic factors included pathological tumour and lymph node stage. | LARC (IOERT n=151; HDR-brachytherapy n=64; data available for 91% of patients) Death No statistically significant difference in in-hospital mortality between the treatments (p=0.546):
Major complications No statistically significant difference in major complications (Clavien-Dindo grade 3 or higher) between the treatments (p=0.665):
Most common major complications were:
LRRC (IOERT n=112; HDR-brachytherapy n=46; data available for 99% of patients) Death No statistically significant difference in in-hospital mortality between the treatments (p>0.999):
Major complications Statistically significantly more major complications (Clavien-Dindo grade 3 or higher) with HDR-brachytherapy than IOERT (p=0.017):
Most common major complications were:
|
Haddock, 2011 Case series | OS 5-year OS: 30% 10-year OS: 16% Significant prognostic factors were treatment era (better survival for more recently treated patients), no prior chemotherapy, and radicality of the resection (R0 better than R1 better than R2). Disease relapse 5-year central relapse (within the IOERT field): 14% 5-year local relapse: 28% 5-year distant relapse: 53% Significant prognostic factors were previous irradiation (previously irradiated patients were more likely to relapse) and radicality of resection (R0 resection better than R1/2) | There were 621 treatment-related complications that affected 302 patients. Of those, 302 complications were considered severe or greater. Deaths 6 patients (<1%) died without relapse of treatment-related complications, and none had been previously irradiated.
IOERT-related severe complications
|
Holman, 2017 Case series | Local re-recurrence 5-year local re-recurrence: 45.3% Significant prognostic factors were preoperative treatment, waiting time between preoperative treatment and surgery, and radicality of the resection. Distant metastases-free survival 3-year distant metastases-free survival: 50% 5-year distant metastases-free survival: 43% Significant prognostic factors were preoperative treatment, radicality of the resection, postoperative radiotherapy. Cancer-specific survival 3-year cancer-specific survival: 62% 5-year cancer-specific survival: 41% Significant prognostic factors were preoperative treatment and radicality of the resection. OS 3-year OS: 52% 5-year OS: 33% Significant prognostic factors were preoperative treatment and radicality of the resection. | Not reported. |
Holman 2016 Case series | Local recurrence 5-year local recurrence: 19.3% Significant prognostic factors were time between preoperative treatment and surgery, and radicality of resection. Distant metastases-free survival 5-year distant metastases-free survival: 64% Significant prognostic factor was radicality of the resection. Cancer-specific survival 5-year cancer-specific survival: 64.6% Significant prognostic factor was radicality of the resection. Relapse-free survival 5-year relapse-free survival: 55.1% Significant prognostic factor was radicality of the resection. OS 3-year OS: 73% 5-year OS: 56% Significant prognostic factors were radicality of the resection and age. | Not reported. |
Kusters, 2010 Case series | Local recurrence 5-year local recurrence rate: 12% Significant prognostic factors were no downstaging, lymph node positivity, margin involvement, and no adjuvant chemotherapy. Distant recurrence 5-year distant metastases recurrence: 29.2% Significant prognostic factors were male gender, preoperatively staged T4 disease, no downstaging, lymph node positivity, and margin involvement. Cancer-specific survival 5-year cancer-specific survival: 73.5% Significant prognostic factors were male gender, lymph node positivity, and margin involvement. OS 5-year OS: 67.1% Significant prognostic factors were age older than 70 years, male gender, no downstaging, lymph node positivity, margin involvement and no adjuvant chemotherapy. | Not reported. |
Procedure technique
In the 2 systematic reviews and meta-analyses, there were 1 and 3 studies that used HDR-brachytherapy, respectively (Fahy, 2021; Liu, 2021). All other studies used IOERT. The devices used to deliver IOERT were not well described. The IOERT dose was typically 10 to 12.5 Gy, with some studies permitting higher doses.
Efficacy
Local recurrence/control outcomes
Seven studies reported local recurrence or local control.
The 2 systematic reviews and meta-analyses reported different findings. In the Fahy (2021) meta-analysis of 7 studies, there was no statistically significant difference between IORT and surgery/EBRT in locoregional recurrence rates, OR 0.55 (95% CI 0.27 to 1.14, p=0.11; figure 1)
 |
In the Liu (2021) meta-analysis of 14 studies, 5-year local control was statistically significantly better with IORT than surgery/EBRT, OR 3.07 (95% CI 1.66 to 5.66, p=0.000, Liu, 2021). When analysing results by study design, this difference was observed in observational studies (OR 3.45; 95% CI 1.54 to 7.73, p=0.000), but not RCTs (OR 1.37; 95% CI 0.35 to 5.35, p=0.655; Figure 2).
 |
The cohort study of IOERT compared with HDR-brachytherapy in people who had an R1 resection found that, after multivariate analysis, people who had IOERT were less likely to experience local recurrence-free survival than people who had HDR-brachytherapy. In people with LARC, in multivariate analysis, local recurrence-free survival was borderline statistically significantly worse with IOERT (HR 0.504; 95% CI 0.254 to 0.999, p=0.050). Similar findings were reported in LRRC, when in multivariate analysis, local recurrence-free survival was statistically significantly worse with IOERT (HR 0.567; 95% CI 0.349 to 0.920, p=0.021; Voogt, 2021).
Two case series reported outcomes for recurrent colorectal cancer or LRRC. In one, 5-year central relapse (within the IOERT field) was 14% and 5-year local relapse was 28% (Haddock, 2011). In the other, 5-year local re-recurrence was 45.3% (Holman, 2017). The common prognostic factor reported by both studies was radicality of the resection.
Two case series reported outcomes for LARC. Five-year local recurrence ranged from 12% (Kusters, 2010) to 19.3% (Holman, 2016). The common prognostic factor reported by both studies was radicality of the resection.
OS
Six studies reported OS.
In a meta-analysis of 9 studies, Liu (2021) found no statistically significant difference in OS between IORT and surgery/EBRT (HR 0.80; 95% CI 0.60 to 1.06, p=0.189; Liu, 2021; figure 3).
 |
The cohort study of IOERT compared to HDR-brachytherapy in people who had an R1 resection reported that there was no statistically significant difference in OS between people who had IOERT and people who had HDR-brachytherapy (LARC, p=0.989; LRRC, p=0.747; Voogt, 2021).
Two case series reported outcomes for recurrent colorectal cancer or LRRC. Five-year OS ranged from 30% (Haddock, 2011) to 33% (Holman, 2017). Haddock (2011) reported 10-year OS of 16%. Common prognostic factors reported by both studies were radicality of the resection and preoperative treatment. Haddock (2011) additionally reported that patients treated after 1997 had better OS.
Two case series reported outcomes for LARC. Five-year OS ranged from 56% (Holman, 2016) to 67.1% (Kusters, 2010). Common prognostic factors reported by both studies were radicality of the resection and age.
Disease-free survival
One study reported disease-free survival.
In a meta-analysis of 6 studies, Liu (2021) found no statistically significant difference between IORT and surgery/EBRT (HR 0.94; 95% CI 0.73 to 1.22, p=0.650; Liu, 2021).
 |
Cancer-specific survival
Three studies reported cancer-specific survival.
One case series reported outcomes for LRRC. Five-year cancer-specific survival was 41% (Holman, 2017). Prognostic factors were preoperative treatment and radicality of the resection.
Two case series reported outcomes for LARC. Five-year cancer-specific survival ranged from 64.6% (Holman, 2016) to 73.5% (Kusters, 2010). The common prognostic factor reported by both studies was radicality of the resection.
Distant metastases
Four studies reported distant metastases outcomes.
Two case series reported outcomes for recurrent colorectal cancer or LRRC. In one study, 5-year distant relapse was 53% (Haddock, 2011). In the other study, 5-year distant metastases-free survival was 43% (Holman, 2017). Common prognostic factors reported by both studies were preoperative treatment and radicality of the resection.
Two case series reported outcomes for LARC. Five-year distant metastases-free survival was 64.6% in Holman (2016) Five-year distant metastases recurrence was 29.2% in Kusters (2010). The common prognostic factor reported by both studies was radicality of the resection.
Safety
Death
One case series reported that for people with recurrent colorectal cancer, 6 patients (less than 1%) died of treatment-related complications without relapse of disease. One death was due to bleeding from the colostomy, and 5 were due to complications arising after surgical correction bowel complications (Haddock, 2011).
One cohort study reported that in people with LARC, there was no statistically significant difference in in-hospital mortality between IOERT and HDR-brachytherapy (HDR-brachytherapy 0%; IOERT 1%; p=0.546). Similarly, in people with LRRC, there was no statistically significant difference in in-hospital mortality between the treatments (HDR-brachytherapy 2%; IOERT 4%, p>0.999; Voogt, 2021).
Severe complications
One cohort study reported that in people with LRRC, there were statistically significantly more major complications (Clavien-Dindo grade 3 or higher) with HDR-brachytherapy than IOERT (HDR-brachytherapy 46%; IOERT 26%, p=0.017; Voogt, 2021). No statistically significant difference was observed in people with LARC (p=0.546).
One case series reported that in people with recurrent colorectal cancer, there were 302 complications considered severe or greater (Haddock, 2011).
Reintervention rate
One meta-analysis compared the reintervention rate between IORT and surgery/EBRT. In the Fahy (2021) meta-analysis of 3 studies, there was no statistically significant difference in reintervention rate between IORT and surgery/EBRT (OR 1.13; 95% CI 0.43 to 2.98, p=0.80).
Specific complications
Wound complications/infections
Two studies compared wound complications or infections between IORT and surgery/EBRT. In the Fahy (2021) meta-analysis of 4 studies there was no statistically significant difference in wound infections between IORT and surgery/EBRT (OR 1.13; 95% CI 0.50 to 2.54, p=0.76). Similarly, in the Liu (2021) meta-analysis of 8 studies, there was no statistically significant difference in wound complications between IORT and surgery/EBRT (OR 1.02; 95% CI 0.52 to 2.02, p=0.948).
In the Voogt (2021) cohort study, abdominal wound dehiscence with evisceration was reported in 11% and 8% of people with LARC and LRRC, respectively, and perineal wound necrosis was reported in 5% of people with LARC. It was not clear how many were observed in people who had IOERT compared against people who had HDR-brachytherapy.
In the Haddock (2011) case series, 7% of people had a wound infection, abscess or fistula that was classified as severe.
Abscess
Two meta-analyses compared abscess or pelvic abscess between IORT and surgery/EBRT. In the Fahy (2021) meta-analysis of 3 studies, there was no statistically significant difference in pelvic abscess rates between IORT and surgery/EBRT (OR 1.01; 95% CI 0.54 to 1.87, p=0.99). Similarly, in the Liu (2021) meta-analysis of 6 studies, there was no statistically significant difference in abscess rates between IORT and surgery/EBRT (OR 1.10; 95% CI 0.67 to 1.80, p=0.833).
In the Voogt (2021) cohort study, presacral abscess was reported in 27% and 26% of people with LARC and LRRC, respectively, and intraabdominal abscess was reported in 9% and 6%, respectively. It was not clear how many were observed in people who had IOERT compared against people who had HDR-brachytherapy.
Anastomotic leak
Two meta-analyses compared anastomotic leak between IORT and surgery/EBRT. In the Fahy (2021) meta-analysis of 5 studies, there was no significant difference in anastomotic leak between IORT and surgery/EBRT (OR 1.06; 95% CI 0.51 to 2.18, p=0.88). Similarly, in the Liu (2021) meta-analysis of 7 studies, there was no statistically significant difference in anastomotic leak between IORT and surgery/EBRT (OR 1.09; 95% CI 0.59 to 2.02, p=0.775).
In the Voogt (2021) cohort study, anastomotic leakage was reported in 5% of people with LARC. It was not clear how many were observed in people who had IOERT compared against people who had HDR-brachytherapy.
Fistulae
One meta-analysis compared fistulae between IORT and surgery/EBRT. In the Liu (2021) meta-analysis of 3 studies, there was no statistically significant difference in fistulae between IORT and surgery/EBRT (OR 0.79; 95% CI 0.33 to 1.89, p=0.600).
In the Haddock (2011) case series, 1% of people had a gastrointestinal tract fistula or obstruction that was classified as severe.
Neurogenic bladder dysfunction
One meta-analysis compared neurogenic bladder dysfunction between IORT and surgery/EBRT. In the Liu (2021) meta-analysis of 3 studies, there was no statistically significant difference between IORT and surgery/EBRT (OR 0.69; 95% CI 0.31 to 1.55, p=0.369).
Neuropathy
In the Haddock (2011) case series, 3% of people had neuropathy that was classified as severe.
Bladder complications
In the Voogt (2021) cohort study, ureter stenosis was reported in 5% of people with LARC. Leakage of the ureter or bladder reconstruction was reported in 5% and 12% of people with LARC and LRRC, respectively. It was not clear how many were observed in people who had IOERT compared against people who had HDR-brachytherapy.
In the Haddock (2011) case series, 3% of people had ureteral obstruction that was classified as severe.
Bleeding
In the Voogt (2021) cohort study, bleeding was reported in 11% of people with LARC. It was not clear how many were observed in people who had IOERT compared to people who had HDR-brachytherapy.
Anecdotal and theoretical adverse events
Expert advice was sought from consultants who have been nominated or ratified by their professional society or royal college. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they have never happened (theoretical). For this procedure, the professional experts did not list any anecdotal or theoretical adverse events.
Five professional expert questionnaires for this procedure were submitted. Find full details of what the professional experts said about the procedure in the specialist advice questionnaires for this procedure.
Validity and generalisability
Seven studies were included in the key evidence summary, including 2 systematic reviews and meta-analyses, 1 cohort study, and 4 case series. Research was done in various countries in Europe and worldwide, though there were no UK studies included in the key evidence.
Both systematic reviews contained comparative studies only.
Fahy (2021) included 2 RCTs. Liu (2021) incorrectly reported that 3 RCTs were included: Masaki (2008) should have been treated as an interim analysis of Masaki (2020). Both RCTs (Dubois, 2011 and Masaki, 2020) were small (fewer than 75 patients per arm). The other studies included in Fahy (2021) and Liu (2021) were observational.
The systematic reviews reported conflicting results: Fahy (2021) found no statistically significant difference in locoregional recurrence between IOERT and surgery/EBRT. Liu (2021) found statistically significantly better local control with IOERT than surgery/EBRT. However, this difference was driven by results from observational studies; a separate analysis of RCTs found no difference.
The authors cite a lack of data concerning the radicality of resection and heterogeneity in patient selection and dosing as potentially confounding factors.
In the other comparative study included (Voogt, 2021), IOERT was associated with statistically significantly worse outcomes than HDR-brachytherapy, after an R1 resection. However, IOERT was associated with statistically significantly fewer severe complications. The authors conclude that further research is required to refine the IOERT procedure.
Given that IOERT is typically used as part of multimodality treatment that may include preoperative (chemo)radiotherapy, resection, IOERT, and postoperative (chemo)radiotherapy, it is difficult to assess the efficacy of IOERT in the non-comparative case series.
The case series show that the most important factor associated with efficacy is the radicality of resection, with R0 resections associated with the best survival and recurrence outcomes.
The multimodality nature of treatment, and the effect size of resection radicality, may mask the true efficacy of IOERT. Both meta-analyses suggest IOERT efficacy, but the authors note that further research is needed to fully understand this.
The majority of patients had LARC or LRRC; there was very limited data on IOERT for colon cancer (all from Haddock, 2011).
The follow-up periods, when reported, ranged from 25 to 80 months.
The authors report no industry funding nor conflicts of interest.
How are you taking part in this consultation?
You will not be able to change how you comment later.
You must be signed in to answer questions