Academiejaar 2015 – 2016

ENDOSCOPISCHE STENTING VAN OBSTRUCTIEF COLONCARCINOOM ALS OVERBRUGGING NAAR ELECTIEVE HEELKUNDE Retrospectieve studie van korte- en langetermijnresultaten in een Belgische cohorte

Annelien VAN DRIESSCHE

Promotor 1: Dr. Vandervoort Jo

Promotor 2: Prof. Dr. De Vos Martine

Masterproef voorgedragen in de master in de specialistische geneeskunde Afstudeerrichting Gastro-Enterologie

Academiejaar 2015 – 2016

ENDOSCOPISCHE STENTING VAN OBSTRUCTIEF COLONCARCINOOM ALS OVERBRUGGING NAAR ELECTIEVE HEELKUNDE Retrospectieve studie van korte- en langetermijnresultaten in een Belgische cohorte

Annelien VAN DRIESSCHE

Promotor 1: Dr. Vandervoort Jo

Promotor 2: Prof. Dr. De Vos Martine

Masterproef voorgedragen in de master in de specialistische geneeskunde Afstudeerrichting Gastro-Enterologie VOORWOORD

Deze masterproef omvat een retrospectieve studie die werd uitgevoerd in het Onze- Lieve-Vrouwziekenhuis te Aalst, waar door de jaren heen ruime ervaring is ontwikkeld in het plaatsen van endoscopische stents in zowel palliatieve als potentieel curatieve setting. De scriptie bevat het manuscript dat op dit moment onder review is bij het tijdschrift ‘Endoscopy’.

Alvorens mijn scriptie aan te vatten wil ik echter graag nog enkele mensen bedanken.

Allereerst gaat mijn dank uit naar Bram. De ideale collega, een goede vriend. Steeds dankbaar voor het verrichte werk, altijd behulpzaam voor wat nog te doen stond. Zeer perfectionistisch in formulering en tekstvorm, liet je soms geen spaander heen van mijn originele schrijfsels. Dank voor je eeuwige enthousiasme en doorzettingsvermogen. Het was fantastisch om met jou samen te werken!

Uiteraard dank aan mijn promotor, Dr. Jo Vandervoort voor het opstarten van deze database en het aanreiken van dit boeiende onderwerp voor mijn masterthesis. Bedankt aan Dr. Jo Vandervoort en aan Dr. Marc De Man om, zonder de kritische blik te verliezen, steeds waardering te tonen voor het geleverde werk. Dankjewel voor de positieve feedback.

Bedankt aan mijn ouders, zus en schoonbroer, voor de morele steun alsook voor het opvangen van de kindjes wanneer wij het druk hadden.

Tenslotte een woord van dank aan mijn prachtige gezinnetje. Mijn lieve echtgenoot, rots in de branding tussen hectische verbouwingen en vele, vele slapeloze nachten door. Mijn schatten van kindjes, die een deel van het opzoekwerk doorbrachten bij mij op schoot of aan de borst. De tijd die nu opnieuw vrijkomt, komt jullie toe. Long-term survival after endoscopic stenting as a bridge to surgery in obstructive colon cancer

Bram Verstockt 1,2 * Annelien Van Driessche 1,3,4 * Marc De Man 1,3 Pieter van der Spek 1 Koen Hendrickx 1 Veerle Casneuf 1,3 Pieter Dobbels 1 Yves Van Molhem 5 Jo Vandervoort 1

1 Dept. Gastroenterology, OLV Hospitals Aalst/Asse/Ninove, Belgium 2 University Hospitals Leuven, Belgium 3 Dept. Gastroenterology, UZ Gent, Belgium 4 Dept. Gastroenterology, AZ Glorieux, Ronse, Belgium 5 Dept. Abdominal Surgery, OLV Hospitals Aalst/Asse/Ninove, Belgium

* co-first authorship !!!

Correspondence author: Jo Vandervoort OLV Hospitals Aalst / Asse / Ninove Moorselbaan 164, B-9300 Aalst, BELGIUM 0032 53 72 44 28 [email protected]

Key words: colon cancer, BTS, SEMS, survival, obstruction

Total word count: 3498 (Introduction, Patients and Methods, Results, Discussion) ABSTRACT

Background and aims: Self-expandable metallic stents are increasingly used in the treatment of obstructive colon cancer (CC). Although endoscopic stenting is widely accepted in a palliative setting, disagreement exists about its role in a curative setting. This study aims to describe long-term survival data in a large patient group, treated with a stent as a bridge to surgery (BTS) for CC.

Patients and methods: Ninety-seven patients, who presented in a Belgian hospital between 1998 and 2013 with a large bowel obstruction due to CC, were included. All patients underwent endoscopic stenting as a BTS in a potentially curable disease. Procedure-related complications and long-term follow-up data were collected and compared with the CC mortality in Belgium in the same era.

Results: Overall survival in this observational cohort did not differ significantly from survival in all Belgian CC patients in the same period (p=0.14). One-, five- and ten- year survival was similar in both groups (95.9% vs 79.0%; 54.7% vs 51.2%; 41.0% vs 35.6% respectively). Technical success rate was 94.8%. Seventy patients did not experience any complication after colonic stenting. Stent migration occurred in 9 patients, whereas micro- and macro perforations were observed in 14 patients, without influencing survival. Incidence rates of peritoneal metastases did not differ between patients with and without any type of perforation (22.2% vs 15.2% respectively, p=0.47).

Conclusion: These data indicate that stenting before surgery is effective and safe in the treatment with curative intent of patients with obstructive CC and reinforce the debate on stenting as a BTS. INTRODUCTION

Colorectal cancer is the third most common cancer in men and the second in women [1]. Approximately 4.5% of men and women will be diagnosed with colorectal cancer at some point during their lifetime [2]. About 10 to 30% of the patients with colorectal cancer present with an acute intestinal obstruction, especially if the tumour is located in the left hemicolon [3,4]. Malignant large bowel obstruction is one of the most common causes of emergency surgery in colorectal cancer. However, urgent surgery is associated with mortality and morbidity rates that are both significantly higher than after elective surgery [3].

In the last decades self-expandable metal stents (SEMS) are increasingly used in the treatment of obstructive colorectal cancer. Endoscopic stenting is widely accepted in a palliative or inoperable setting [5], whereas disagreement exists about its role as a standard therapy in a curative setting [6,7].

Nevertheless, there has been growing interest in using SEMS for bowel decompression as a bridge to surgery (BTS) for patients with curable disease. The main advantage of this approach is the elective, and thereby less invasive, character of the surgical resection [8]. It offers the opportunity for an adequate pre-operative assessment of the patient, correction of dehydration and electrolyte disturbances and a good preparation of the colon. Subsequently, reported advantages are a decreased anaesthetic risk, a shorter hospitalisation and intensive care unit stay, lower rate of (sub)total colectomy and necessity for stoma [9]. In complicated urgent surgery, the oncologic outcome can be negatively affected by delayed or omitted adjuvant chemotherapy [10], due to the difficult wound healing and general recovery or the need for additional surgery [9,11,12].

Contrary, SEMS placement has its own potential risks, such as bowel perforation (due to the stent itself or to air insufflation during the procedure), stent migration, bleeding and clinical failure (absence of transit or reobstruction). Furthermore, technical issues might occur.

Finally, there exists a specific oncologic concern in the curative setting about the risk of tumour dissemination, caused by stent-induced dilatation or perforation, which might negatively affect the long-term survival [13-16]. So far, few data are available on the long-term outcome of SEMS in patients with an obstructive, but potentially curable colon cancer (CC) [7,8,14,16-19]. This study aims to provide success rates, complications, surgical features and long-term survival data on the use of SEMS in obstructive CC as a bridge to surgery.

PATIENTS AND METHODS

In this retrospective study, we included all patients who presented at the OLV Hospital Aalst (Belgium) between June 1998 and November 2013, with a large bowel obstruction, due to CC, and who underwent endoscopic stenting as a BTS in a potentially curable disease. All patients gave informed consent before the endoscopic procedure. Prior to any therapeutical intervention, radiologic staging was performed at diagnosis of the CC.

Two experienced interventional gastroenterologists (JVDV and PVDS) performed all endoscopic stenting procedures under combined endoscopic/fluoroscopic guidance. They used two types of SEMS: Wallstent Enteral and WallFlex colonic stent, both manufactured by Boston Scientific.

Technical success was defined as a successful stent placement at the end of the endoscopic procedure, radiologically confirmed. Clinical success was defined as a clinical and radiological evidence of colonic decompression within 72 hours after stent insertion.

If no per or post procedural complications, patients were often discharged, followed by surgery a couple of weeks later. Time to surgery mainly depended on the global patient condition, complications and the recovery of the colonic distension. Abdominal surgery was performed at the same centre. Surgical technique used was decided upon surgeons’ preference. If possible, laparoscopy was preferred to laparotomy.

The need for any adjuvant therapy (or additional metastasectomy when a metastasis was found perioperative), was decided at multidisciplinary oncological meetings.

Baseline and long-term outcome data (up to December 2015) were obtained from the corresponding medical records. If necessary, additional information was received via the patient’s general practitioner.

Statistical analyses were performed using R (http://www.r-project.org/). Survival rates were calculated using the Kaplan-Meier method. To compare differences between survival curves, Mantel-Haenzel test was used. A Chi-square test was used to compare proportions. A p-value <0.05 was considered as being statistically significant. RESULTS

Patient characteristics

Of all 342 colonic stenting procedures performed during the study period, 114 were started aiming a BTS. Eleven patients were excluded from the overall analysis because histological review of the resection specimens revealed another etiology of the intestinal narrowing, although CC was initially expected based on radiologic and endoscopic findings. Three patients did not have surgery after SEMS placement and were therefore excluded from the overall analysis. Finally 3 patients were excluded from long-term outcome analysis because of incomplete tumour staging or loss of follow-up. In total 97 patients were enrolled in the overall analysis.

Baseline characteristics of these 97 patients are listed in table 1. The data cover a total follow-up period of 501.7 patients-years (mean follow-up 5.2 years, range 0.2– 14.7 years).

All patients presented initially with a large bowel obstruction, caused by a CC (96 adenocarcinoma, 1 signet-ring carcinoma). In 1 patient a bifocal tumour was found in the rectosigmoid. Among all included patients the Tumour Node Metastasis (TNM) staging is rather heterogeneous, with a predominant presence of stage IIa and IIIb tumours. Metastatic disease at diagnosis was found in 17.5% of the patients, based on the perioperative findings and additional imaging after colonic stenting (Table 2).

Additionally to the endoscopic and surgical procedure, 63 patients received further treatment, based on the current international guidelines for CC, after a multidisciplinary oncologic discussion. In 34 patients no chemotherapy was administered either because there was no indication (16 patients, 16.5%) or because of age-related comorbidities (15 patients, 15.4%) or refusal of patient (3 patients, 3.1%) (Table 2).

Technical/clinical success rates of endoscopic stenting

The endoscopic procedure could restore the colonic transit in 93.8% (n=91) of all patients.

Technical success rate was 94.8%. In 5 patients the stenting procedure was not successful due to technical problems (Table 3): in 4 there was an inability to pass the guidewire and in 1 a technical defect to the colonoscope (rupture of cable) occurred. In 3 of these 5 patients, surgery was performed on the same day. In 1 patient an urgent decompressive caecostomy was created and an elective partial colectomy was carried out 2 weeks later. One patient was managed conservatively after stenting failure until elective surgery 5 days later. We recorded clinical failure in 1 patient with no production of flatus or stool despite technically successful stenting procedure. Computed-tomography revealed an important faecal impaction. A decompressive caecostomy was performed 2 days after the stenting procedure, followed by a rectosigmoidectomy in a second time 6 days later.

Complications post stenting

Seventy patients (72.2%) did not experience any complication after the colonic stenting (Table 3). Stent migration occurred in 9 patients. Depending on the clinical consequences (induced obstruction or not), surgery was performed on the same day or later on (range 0 – 14 days). In 1 patient a second stent was placed after migration of the first in order to restore transit. In 2 patients stent migration occurred unnoticed: in 1 the dislocated stent was discovered peroperatively; the second evacuated the stent unnoticed.

We observed any type of stent perforation (micro or macro) in 14 patients. No perprocedural perforations were observed. One patient developed a caecal blow-out a few hours after stenting procedure, necessitating immediate surgery. Two patients presented at least 1 week after the procedure with symptoms of abdominal pain (up to acute abdomen), caused by stent migration to the proximal colon above the tumour, which induced perforation and compelled urgent surgery. In 9 patients there were no clinical signs pre-operatively, but perforation was observed macroscopically peroperatively. Finally 2 cases of micro perforation (without any notice of macro perforation peroperatively) were seen during histologic examination of the resection specimen.

In 2 patients a caecum perforation occurred at least a few days after endoscopic stent placement: in 1 patient the perforation was suspected to be in the caecum based on the CT-graphical distribution of intraperitoneal air and fluid, although the exact location could not be proved radiologically. The perforation itself was managed conservatively with antibiotics and surgery took place 10 days later. In a second patient the caecal perforation developed 4 days after stenting and seemed driven by ischaemia, based on pathological analysis of the resection specimen. Though, no arguments for reobstruction or caecal blow-out existed after the stenting procedure in this particular patient. In 1 additional case the colon was initially intact but the stent perforated the colonic wall as a consequence of manipulation by the surgeon. Furthermore, in 1 patient the tumour itself infiltrated transmurally, causing a clinical clear perforation compelling immediate surgery.

Finally, 1 patient developed an E. Coli sepsis as a result of stent dysfunction with secondary obstruction. Peroperatively a suppurative peritonitis was identified, although no evident perforation could be found (neither macroscopically by the surgeon nor by histologic examination of the resection specimen). Surgical procedure

Sixty point two per cent of the patients were discharged home 6.8 (mean) days after the endoscopic procedure. Overall, surgery took place 16.6 (mean) days after colonic stenting, ranging from an operation on the same day as the endoscopic procedure, to an interval of maximal 124 days. Patients who stayed in hospital between both procedures were operated after 11.0 (mean) days, whereas patients who went home in between had surgery after 21.3 days on average.

In 85.6% of the cases a laparoscopic resection of the obstructive tumour was performed (Table 4). Nine point three per cent of the patients got primarily open surgery. In 5.2% of the patients a laparoscopic procedure was converted to laparotomy, because of adhesions or peritonitis. The performed procedure depended on the location of the tumour and the peroperative findings (Table 4).

Per- and postoperative complications

One splenectomy was performed after iatrogenic spleen laceration during surgery. Surgical revision was necessary in 7 patients (7.2%) because of one or more of the following complications: bleeding (n=2), intestinal ischaemia/necrosis (n=3), abcedation (n=2), anastomotic leakage (n=2).

One patient died as a result of sepsis after several postoperative complications: necrotic anastomosis after a laparascopic anterior resection, for which a Hartman procedure was performed; subsequent wound infection with abcedation and loosening of the stoma; sepsis eventually leading to exitus.

Other complications observed during the first postoperative days were catheter sepsis (n=1), pancreatitis (n=2) and urinary tract infection (n=1).

Survival and outcome

The 1-, 3-, 5- and 10-year survival according to different stages of CC (American Joint Committee on Cancer, 7th Edition) are presented in Table 5 and Supplementary Table 1. The Belgian Cancer Registry (BCR) data (2004-2013) for CC are additionally added to place our results in the right perspective [20]. We need to point out that the majority of CC patients in this last group did not present with obstructive disease. Comparing the overall survival (OS) rate in our observational cohort with the OS of CC in all Belgian patients in approximately the same era (Figure 1), shows no statistical difference (p=0.14). Even so, for stages which we could include at least 10 patients (i.e. stage II, III and IV), we did not observe any statistical difference with the data of the BCR (p=0.21, p=0.58, p=0.10 respectively).

Because of the oncologic concern of endoscopic stenting (a perforation with an associated possibility of peritoneal metastasis might occur), an additional survival sub-analysis was made of all 14 patients who had had a stent-related perforation. Oncologic staging of this small group was rather heterogeneous (4 stage IIa, 5 IIIb, 2 IIIc, 3 IV). Only the overall 1- (100.0%), 3- (84.4%) and 5-year (54.3%) survival is known, 10-year survival is not available. Comparing survival rates of these patients (n=14) with those who had no stent-related perforation (n=83), did not reveal any difference (p=0.72) (Figure 2). Likewise, no difference was observed between the OS rates of these 14 perforated patients and the OS rates reported by the BCR (p = 0.43).

Finally, 4 of the 18 patients (22.2%) with any type of perforation developed peritoneal metastases later on, which did not significantly differ (p=0.47) from the proportion of patients without any perforation (12 out of 79; 15.2%) who developed peritoneal metastases later on in the disease.

DISCUSSION We report, to our knowledge, the longest follow-up period ever described in a large patient group treated with a stent as a BTS for obstructive, but potentially curable CC.

All our data about survival rates, safety and efficacy support our statement that colonic stenting in obstructive CC still has a place in the curative setting (Figure 1, Table 1 and Supplementary Table 1, Figures 1, 2 and 3, Table 3). The experience of the operating endoscopist is probably one of the key factors in achieving acceptable survival rates, before it can be safely implemented in a digestive oncologic care program. Our reported overall 5-year survival (54.7%) approximates rates reported by other groups in patients with obstructive CC [8,13,16,17].

To date, there are very few studies reporting on long-term survival after stent insertion as a BTS in a potentially curative setting. Iversen et al. reported retrospectively a 3-year survival of 74% after SEMS insertion as a BTS [21]. Equally, a 3-year survival of 80%, not significantly different from survival rates after surgery (74%), was observed by Dastur [22]. Though, SEMS insertion negatively influenced 5-year survival and disease-free survival (DFS) in stage II and III (5-year survival rate of 44% after SEMS versus 87% after elective surgery for non-obstructing CC) in a sub-analysis by Kim et al [14]. Nevertheless it is unclear what magnitude of this effect is related to the underlying obstruction rather than to the stenting. Similarly, Sabbagh et al. published a worse 5-year survival after SEMS (25%) compared to immediate surgery (62%), though there were no significant differences in terms of 5-year DFS [18].

Contrary, Knight et al. concluded that stent insertion as BTS did not have any adverse effect on long-term survival (5-year survival of 60 % after SEMS versus 58 % after elective surgery for non-obstructive colorectal cancer) [17]. Tung et al. also found no significant difference in the 5-year survival when comparing this endo- laparoscopic approach with conventional open surgery (42.8% versus 57.1%, p=0.347) in patients with stage II and III [8]. Other groups also published comparable 5-year survival rates between both SEMS treated and conventionally treated patients [16,19]. Moreover, Gianotti et al. reported a benefit for SEMS during a 3-year follow- up period [23]. Finally, in a recent meta-analysis of 11 studies, including 1136 patients, BTS was oncologically comparable to emergency surgery with respect to OS, DFS and recurrence [7].

Generally, our overall 1-, 3-, 5- and 10-year survival across all tumour stages is not worse compared to the survival rates reported in most of the previous cited studies.

Both our technical and clinical success rates compare favourably with the pooled success rates of a recent meta-analysis including 405 BTS patients [24]. The favourable outcomes in the current study can be explained by the availability of two experienced endoscopists (JVDV and PVDS) who regularly perform stenting procedures. Although many of the included patients did not experience any complication at all, migration occurred in 9.3% of the cases, comparable to previous reported data [24,25].

To evaluate the rate of perforations, we not only included those with clinical repercussions, but also minor perforations seen peroperatively and micro- perforations found upon histologic analysis of the resection specimen. This explains a rather high perforation rate (14.4%) in our analysis, similar to other groups [24,26].

The oncological consequences of potential tumour dissemination, caused by perforations, are unclear, but the possibility of dissemination is worrisome and these silent perforations should not be disregarded [26,27]. Nevertheless, the survival of patients with any stent-related perforation did not differ statistically from the survival of those who did not have a perforation (Figure 2). Although both groups are rather small and thus potentially underpowered to detect small significant differences, this real-life observation might be important to reduce the oncologic concern about potential tumour dissemination after stent perforation in the frail and aged patient. To exclude a confounding effect of adjuvant therapy to the reported survival rates in both groups, proportions of adjuvant therapy were compared and did not show any significant difference (data not shown). Moreover, incidence of peritoneal metastases did not differ between groups with or without any type of perforation (p=0.47). Additionally, the OS of patients with stent-related perforations did not differ from the OS reported by the BCR (p=0.43). Importantly, most (71.4%) of these patients with any stent-related perforation had at least stage III disease, whereas in the BCR this was only the case in 42.4% of patients [20], reflecting that this perforated subgroup a priori had a worse prognosis compared to the overall outcome in the comparison cohort. Furthermore, suffering from an obstructive CC, decreased their prognosis even more in advance [28]. But, as their survival does not seem to be different during long-term follow-up, this confirms the safety profile of SEMS, even if a perforation would occur.

After successful stenting, surgery took place after 16.6 days on average. In most publications no criteria for the assessment of the timing of surgery are described. Timing depends on personal habits of the treating physician, which might explain the broad range (2-233 days) found in literature [13,17,19,21,23,24,27,29-32]. The main concern about this timing is the likelihood of reobstruction, caused by migration or obliteration of the stent [31]. In addition, many patients with obstructive CC have advanced disease and thus require adjuvant chemotherapy. Therefore prolonging the duration between stenting and subsequent surgery in this setting may lead to a delay in adjuvant chemotherapy [11,31], which might negatively influence survival [10]. A retrospective study recently suggested that 1 week after stenting should be enough for decompression to prepare for subsequent surgery. Extending this duration exposes the patient to the risk of reobstruction with endoscopic reintervention or emergency surgery, which is associated with a decreased primary anastomosis rate and increased stoma creation [31].

In the majority of patients in our cohort surgery was performed via laparoscopic approach (Table 4). Our stoma rate is low (5.2 %), which can be attributed to the elective character of the surgery. Though, the randomized controlled trial (RCT) by van Hooft et al. reported a stoma rate of 57.4% at latest follow-up [26], in accordance with stoma rates published by other groups [33]. Urgent surgery for obstructive CC traditionally consists of a multistage procedure, with resection and stoma formation, followed by later reversal. However, in about 40% of patients subsequent reconstruction is never performed, because of advanced age or comorbidities [24,34]. The presence of a stoma is known to impact negatively on patients’ quality of life. In those who undergo reversal, this surgery involves additional per- and postoperative risks.

In several recent meta-analyses comparing emergency surgery with elective surgery after stenting, the latter is shown to be associated with a lower risk of stoma creation, higher rate of primary anastomosis, shorter hospital and intensive care unit stays, and lower overall complication rates [3,6,9,24,29,30,33]. Our low stoma rate reinforces this potential benefit of SEMS as BTS, but based on the design of this study we cannot compare with emergency surgery. In most datasets, colon and rectal cancer are described separately (although survival results are almost identical for similar TNM stages). The only included rectal cancers in this study were located on the rectosigmoid junction, and are thus considered to have the characteristics of CC. Furthermore, the absence of a control arm of patients undergoing primary surgery, prevents us to compare survival between these two therapeutic options. Therefore we chose to compare with the survival rates of Belgian CC in the same period, including many patients conventionally treated. Although we realize that this comparison is not ideal and fully matched, we are convinced that the publication of this large, real-life SEMS survival dataset might have an important impact on the debate about its role in obstructive CC.

Despite these limitations, this is one of the largest studies so far concerning long- term survival after stenting as a BTS for obstructive CC in a curative setting. Although current ESGE-guidelines do not advise to use SEMS in potentially curable CC in patients with low operative risk (age <70 and ASA I-II),[5] there is certainly an indication for this technique in the majority of patients presenting with obstructive CC, namely in all palliative settings and in those patients with curable disease but high operative risk, i.e. age >70 and ASA score III-IV. Our data reinforce the debate about a place for stenting as a bridge to surgery in experienced hands, alternatively to urgent surgery, in a curative setting for patients suffering from obstructive CC. Nowadays, metastatic disease not always implies incurable disease. We believe SEMS has its role in those patients with (borderline) resectable metastases, because starting chemotherapy as soon as possible has a higher priority, compared to the potential risk of tumour spreading by stenting. Because of the existing, conflicting literature, we are looking forward to the results of new RCTs focussing on the survival after stenting in comparison with primary resection. CONFLICTS OF INTEREST: All authors declare no conflicts of interest or financial relationships.

ACKNOWLEDGMENTS: We would like to thank the Belgian Cancer Registry for providing us all necessary data about the long-term survival of colon cancer in Belgium.

RESPONSABILITIES:

BV and AVD did all (statistical) analysis and wrote the manuscript (shared first authorship). JVDV started this dataset. PVDS, MDM, KH, VC, PD, YVM and JVDV carefully reviewed the manuscript before submission. All authors agreed with the final version before submission. FIGURES & FIGURE LEGEND:

Figure 1: Comparison between overall survival of our BTS-patients and the Belgian Cancer Registry data (all stages included). Diamonds clearly indicate the 1-, 3-, 5- and 10- year survival rates respectively. Figure 2: Comparing survival rates of 14 patients with a stent-related perforation and 83 with no stent-related (micro and macro) perforation at all. TABLES AND CORRESPONDING LEGEND:

Table 1. Baseline characteristics of the 97 included patients

Sex, no of female : male 64 : 33

Age, mean (SD) years 70.6 (11.4)

Comorbidity, n (%)

Cardiovascular disease 39 (40.2)

Pulmonary disease 19 (19.6)

Diabetes mellitus 16 (16.5)

Neurologic disease 10 (10.3)

Auto-immune disease 6 (6.2)

Other malignancy 9 (9.3)

HNPCC 1 (1.0)

IBD 1 (1.0) Table 2. Oncological characteristics of the 97 included patients

Tumour localisation, n (%) Rectosigmoid 53 (54.6)

Descending colon 31 (32.0)

Transverse colon 11 (11.3)

Ascending colon 2 (2.1)

TNM Staging, n (%) Stage I 2 (2.1)

Stage II 36 (37.1)

Stage IIa 28 (28.9)

Stage IIb 8 (8.2)

Stage III 42 (43.3)

Stage IIIa 2 (2.1)

Stage IIIb 27 (27.8)

Stage IIIc 13 (13.4)

Stage IV 17 (17.5)

Additional treatment, n (%) Neoadjuvant radio chemotherapy 2 (2.1) between endoscopic stenting and surgery

Adjuvant chemotherapy 48 (49.4)

Adjuvant radio chemotherapy 1 (1.0)

Palliative chemotherapy 13 (13.4)

No additional treatment 34 (5.0) Table 3. Failure rates and complications of stenting procedure

Failure rates of stenting, n (%)

Technical failure 5 (5.2)

Clinical failure 1 (1.0)

Reobstruction after initially successful stenting 3 (3.1)

Complications, n (%)

Stent migration 9 (9.3)

compelling immediate surgery 3

elective surgery next days/weeks 6

Stent-related perforations 14 (14.4)

Symptomatic 2

Asymptomatic - observed peroperatively 9

Asymptomatic – micro perforation (observed 2

during histologic analysis of resection specimen)

Caecal perforation immediately post-stenting 1 (caecal blow-out)

Other perforations 4 (4.1)

1 Tumour induced perforation

1 Clinical suspicion of caecal perforation, treated conservatively (no radiological argument for stent perforation)

1 Caecal perforation 4 days after stenting without reobstruction

Perforation induced during surgery while 1 manipulating the colon Table 4. Characteristics of the surgical procedures

Abdominal access, n (%)

Laparoscopy 83 (85.6)

Laparotomy 9 (9.3)

Conversion from laparoscopy  laparotomy 5 (5.2 )

Stoma, n (%) 5 (5.2)

Temporarily 1

Permanent 4

Performed procedure, n (%)

Low anterior resection 9 (9.3)

High anterior resection 45 (46.4)

Left hemicolectomy 33 (34.0)

Right hemicolectomy 5 (5.2)

Subtotal colectomy 3 (3.1)

Segmental resection 2 (2.0) Table 5. Long-term survival for different CC stages. In bold our observational data. In Italic survival rates of CC based on Belgian Cancer Registry (2003-2014) [20]. NA means we did not have any patients to tell anything about the survival rate at this time point.

All stages STAGE I STAGE II STAGE III STAGE IV

1-year 95.9% 100% 100% 92.9% 94.1% survival 79.0% 92.0% 88.5% 82.5% 57.9%

3-year 65.8% 50.0% 79.8% 67.6% 32.7% survival 61.2% 83.1% 76.4% 63.0% 24.2%

5-year 54.7% 50.0% 76.6% 49.0% 19.6% survival 51.2% 73.5% 66.1% 51.2% 13.9%

10-year 41.0% NA 55.6% 39.9% 13.1% survival 35.6% 52.5% 46.1% 32.8% 7.7% SUPPLEMENTARY FILES

Supplementary Figure 1: Comparison between overall survival of our BTS-patients and the Belgian Cancer Registry data (stage II). Diamonds clearly indicate the 1-, 3-, 5- and 10-year survival rates respectively. Supplementary Figure 2: Comparison between overall survival of our BTS-patients and the Belgian Cancer Registry data (stage III). Diamonds clearly indicate the 1-, 3-, 5- and 10-year survival rates respectively. Supplementary Figure 3: Comparison between overall survival of our BTS-patients and the Belgian Cancer Registry data (stage IV). Diamonds clearly indicate the 1-, 3-, 5- and 10-year survival rates respectively. Supplementary Table 1: Long-term survival for different CC sub-stages. NA means we didn’t have any patients to tell anything about the survival rate at this time point.

STAGE IIa STAGE IIb STAGE IIIa STAGE IIIb STAGE IIIc

1-year 100% 100% 100% 96.3% 84.6% survival

3-year 81.5% 87.5% 0.0% 88.9% 30.8% survival

5-year 81.5% 62.5% NA 63.8% 23.1% survival

10-year 66.4% 20.8% NA 56.7% NA survival REFERENCES

1. Society AC. Global Cancer. Facts & Figures 2008 [Available from: http://www.cancer.org/acs/groups/content/@epidemiologysurveilance/documents/doc ument/acspc-027766.pdf. Last consulted 3rd of January 2016]. 2. SEER Stat Fact Sheets: Colon and Rectum Cancer [Available from: http://seer.cancer.gov/statfacts/html/colorect.html. Last consulted 3rd of January 2016]. 3. Huang X, Lv B, Zhang S et al. Preoperative colonic stents versus emergency surgery for acute left-sided malignant colonic obstruction: a meta-analysis. J Gastrointest Surg 2014; 18: 584-591 4. Kaplan J, Strongin A, Adler DG et al. Enteral stents for the management of malignant colorectal obstruction. World J Gastroenterol 2014; 20: 13239-13245 5. van Hooft JE, van Halsema EE, Vanbiervliet G et al. Self-expandable metal stents for obstructing colonic and extracolonic cancer: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2014; 46: 990-1053 6. Sagar J. Colorectal stents for the management of malignant colonic obstructions. Cochrane Database Syst Rev 2011: CD007378 7. Matsuda A, Miyashita M, Matsumoto S et al. Comparison of long-term outcomes of colonic stent as "bridge to surgery" and emergency surgery for malignant large-bowel obstruction: a meta-analysis. Ann Surg Oncol 2015; 22: 497-504 8. Tung KL, Cheung HY, Ng LW et al. Endo-laparoscopic approach versus conventional open surgery in the treatment of obstructing left-sided colon cancer: long-term follow- up of a randomized trial. Asian J Endosc Surg 2013; 6: 78-81 9. Lee JM, Byeon JS. Colorectal Stents: Current Status. Clin Endosc 2015; 48: 194-200 10. Biagi JJ, Raphael MJ, Mackillop WJ et al. Association between time to initiation of adjuvant chemotherapy and survival in colorectal cancer: a systematic review and meta-analysis. JAMA 2011; 305: 2335-2342 11. Ho KS, Quah HM, Lim JF et al. Endoscopic stenting and elective surgery versus emergency surgery for left-sided malignant colonic obstruction: a prospective randomized trial. Int J Colorectal Dis 2012; 27: 355-362 12. Ghazal AH, El-Shazly WG, Bessa SS et al. Colonic endolumenal stenting devices and elective surgery versus emergency subtotal/total colectomy in the management of malignant obstructed left colon carcinoma. J Gastrointest Surg 2013; 17: 1123- 1129 13. Alcántara M, Serra-Aracil X, Falcó J et al. Prospective, controlled, randomized study of intraoperative colonic lavage versus stent placement in obstructive left-sided colonic cancer. World J Surg 2011; 35: 1904-1910 14. Kim JS, Hur H, Min BS et al. Oncologic outcomes of self-expanding metallic stent insertion as a bridge to surgery in the management of left-sided colon cancer obstruction: comparison with nonobstructing elective surgery. World J Surg 2009; 33: 1281-1286 15. Maruthachalam K, Lash GE, Shenton BK et al. Tumour cell dissemination following endoscopic stent insertion. Br J Surg 2007; 94: 1151-1154 16. Kim HJ, Choi GS, Park JS et al. Higher rate of perineural invasion in stent- laparoscopic approach in comparison to emergent open resection for obstructing left- sided colon cancer. Int J Colorectal Dis 2013; 28: 407-414 17. Knight AL, Trompetas V, Saunders MP et al. Does stenting of left-sided colorectal cancer as a "bridge to surgery" adversely affect oncological outcomes? A comparison with non-obstructing elective left-sided colonic resections. Int J Colorectal Dis 2012; 27: 1509-1514 18. Sabbagh C, Browet F, Diouf M et al. Is stenting as "a bridge to surgery" an oncologically safe strategy for the management of acute, left-sided, malignant, colonic obstruction? A comparative study with a propensity score analysis. Ann Surg 2013; 258: 107-115 19. Choi JM, Lee C, Han YM et al. Long-term oncologic outcomes of endoscopic stenting as a bridge to surgery for malignant colonic obstruction: comparison with emergency surgery. Surg Endosc 2014; 28: 2649-2655 20. Belgian Cancer Registry. Cancer in Belgium, 2013. Brussels. 21. Iversen LH, Kratmann M, Bøje M et al. Self-expanding metallic stents as bridge to surgery in obstructing colorectal cancer. Br J Surg 2011; 98: 275-281 22. Dastur JK, Forshaw MJ, Modarai B et al. Comparison of short-and long-term outcomes following either insertion of self-expanding metallic stents or emergency surgery in malignant large bowel obstruction. Tech Coloproctol 2008; 12: 51-55 23. Gianotti L, Tamini N, Nespoli L et al. A prospective evaluation of short-term and long- term results from colonic stenting for palliation or as a bridge to elective operation versus immediate surgery for large-bowel obstruction. Surg Endosc 2013; 27: 832- 842 24. De Ceglie A, Filiberti R, Baron TH et al. A meta-analysis of endoscopic stenting as bridge to surgery versus emergency surgery for left-sided colorectal cancer obstruction. Crit Rev Oncol Hematol 2013; 88: 387-403 25. Sebastian S, Johnston S, Geoghegan T et al. Pooled analysis of the efficacy and safety of self-expanding metal stenting in malignant colorectal obstruction. Am J Gastroenterol 2004; 99: 2051-2057 26. van Hooft JE, Bemelman WA, Oldenburg B et al. Colonic stenting versus emergency surgery for acute left-sided malignant colonic obstruction: a multicentre randomised trial. Lancet Oncol 2011; 12: 344-352 27. Pirlet IA, Slim K, Kwiatkowski F et al. Emergency preoperative stenting versus surgery for acute left-sided malignant colonic obstruction: a multicenter randomized controlled trial. Surg Endosc 2011; 25: 1814-1821 28. Zucchetti F, Negro F, Matera D et al. Colorectal cancer: obstruction is an independent negative prognostic factor after radical resection. Ann Ital Chir 2002; 73: 421-425 29. Tilney HS, Lovegrove RE, Purkayastha S et al. Comparison of colonic stenting and open surgery for malignant large bowel obstruction. Surg Endosc 2007; 21: 225-233 30. Watt AM, Faragher IG, Griffin TT et al. Self-expanding metallic stents for relieving malignant colorectal obstruction: a systematic review. Ann Surg 2007; 246: 24-30 31. Abdussamet Bozkurt M, Gonenc M, Kapan S et al. Colonic stent as bridge to surgery in patients with obstructive left-sided colon cancer. JSLS 2014; 18: 32. Bayraktar B, Ozemir IA, Kefeli U et al. Colorectal stenting for palliation and as a bridge to surgery: A 5-year follow-up study. World J Gastroenterol 2015; 21: 9373- 9379 33. Cirocchi R, Farinella E, Trastulli S et al. Safety and efficacy of endoscopic colonic stenting as a bridge to surgery in the management of intestinal obstruction due to left colon and rectal cancer: a systematic review and meta-analysis. Surg Oncol 2013; 22: 14-21 34. Frago R, Ramirez E, Millan M et al. Current management of acute malignant large bowel obstruction: a systematic review. Am J Surg 2014; 207: 127-138 KORTE SAMENVATTING MASTERPROEF

ACHTERGROND

Wanneer een patiënt zich presenteert met een obstructie op basis van een coloncarcinoom, is urgente decompressie noodzakelijk. Dit kan door middel van heelkundige resectie of door endoscopische plaatsing van een stent. In een palliatieve of inoperabele setting is stentplaatsing algemeen aanvaard als de standaard behandeling om de obstructie op te heffen. In een potentieel curatieve setting wordt in de huidige richtlijnen primaire heelkunde aangeraden, met resectie van de obstructieve tumor. Er is echter toenemende interesse voor endoscopische stenting in de acute fase ter decompressie, gevolgd door een resectie in tweede tijd. Dit biedt het voordeel dat de heelkunde electief kan gepland worden, met voorafgaande voorbereiding van de patiënt en planning van een optimale operatieve setting. Als gevolg hiervan kan de ingreep op een veiliger en minder invasieve manier gebeuren, met lager anesthesierisico, minder complicaties, en kortere hospitalisatieduur tot gevolg. Doordat de ingreep gebeurt op een relatief gezond colon (buiten acute obstructie) kan het risico op uitgebreide resectie en op plaatsing van een stoma gereduceerd worden. De oncologische uitkomst kan gecompromitteerd worden bij urgente heelkunde in suboptimale omstandigheden, doordat adjuvante chemotherapie moet worden uitgesteld door trager herstel van de patiënt, moeizame wondheling, of de nood aan bijkomende operaties. Anderzijds bestaat in een curatieve setting de vrees voor tumorale uitzaaiing, dewelke zou kunnen in de hand gewerkt worden door druk van de stent of door perforatie van het colon. Tenslotte moet ook worden rekening gehouden met de risico’s van de stentplaatsing op zich, zoals perforatie, bloeding, migratie.

DOELSTELLING

Het doel van deze studie is enerzijds de korte termijnresultaten en complicaties te beschrijven van endoscopische stenting gevolgd door electieve heelkunde; en anderzijds de langetermijnresultaten na te gaan op oncologisch vlak en op algemene overleving.

METHODOLOGIE

In het Onze-Lieve-Vrouwziekenhuis in Aalst werden tussen juni 1998 en november 2013 in totaal 342 stents geplaatst voor opheffen van een obstructie door een coloncarcinoom. In deze retrospectieve studie werden 97 patiënten geïncludeerd bij wie endoscopisch een colonstent werd geplaatst als overbrugging naar latere electieve heelkunde. Gegevens werden verzameld over klinisch en technisch succes van de stentplaatsing, de chirurgische techniek, en eventuele complicaties van de endoscopische en de chirurgische procedures. Bovendien werden lange termijngegevens verzameld over de oncologische evolutie (zoals het ontwikkelen van peritoneale metastasen) en de algemene overleving. Deze data werden gecollecteerd op basis van het medisch dossier in het OLV-ziekenhuis en via contact met huisartsen en met specialisten in andere ziekenhuizen. Tenslotte werden onze resultaten vergeleken met de mortaliteit van Belgische patiënten met coloncarcinoom in dezelfde periode, op basis van data van het Belgisch Kankerregister.

RESULTATEN

Bij 94.8% van de patiënten kon de endoscopische behandeling technisch succesvol worden uitgevoerd, met correcte positie van de stent aan het einde van de procedure. Technisch falen was in de meerderheid van de gevallen (4 van de 5 patiënten) het gevolg van uitgesproken tumorale vernauwing van het lumen, waardoor de voerdraad niet veilig kon worden opgeschoven. Bij 1 patiënt werd de endoscopische procedure vroegtijdig beëindigd wegens een defecte coloscoop. Het klinisch succes was 99.0%: slechts 1 patiënt kwam niet op transit binnen de 72 uur na plaatsing van de stent. 72.2% van de patiënten kende een ongecompliceerd verloop na de stentplaatsing. Er trad 1 caecale blow-out perforatie op vlak na de endoscopische procedure. Bij 9 patiënten was er een spontane migratie van de colonstent. Bij nazicht van klinische gegevens, operatieverslagen en APO-resultaten werden 13 perforaties weerhouden door de stent zelf. Hieronder waren er slechts 2 patiënten symptomatisch, bij 11 patiënten was er geen duidelijk klinisch beeld maar werd een perforatie van de stent door de darmwand vastgesteld peroperatief (9 patiënten) of bij onderzoek van het resectiestuk nadien (2 patiënten).

De chirurgische resectie kon in 83 patiënten via een laparoscopische toegangsweg gebeuren, 9 patiënten kregen een open ingreep, en in 5 gevallen was er conversie van laparoscopie naar laparotomie omwille van verklevingen of peritonitis. Er werden 5 stoma’s geplaatst, waarvan de meeste permanent. Slechts 1 stoma werd in tweede tijd terug afgebroken. Bij 7 patiënten was een chirurgische revisie nodig omwille van postoperatieve complicaties zoals bloeding, intestinale ischemie, abcedatie of anastomoselek. 1 patiënt overleed in sepsis na een ernstige gecompliceerd postoperatief verloop.

De algemene overleving in onze cohorte bedroeg 95.9% na 1 jaar, 65.8% na 3 jaar, 54.7% na 5 jaar en 41.0% na 10 jaar. Er is geen statistisch verschil te weerhouden (p=0.14) bij vergelijking met de algemene overleving van darmkankerpatiënten in België. Ook de overleving per tumorstadium verschilt niet significant met deze gerapporteerd door het Belgisch Kankerregister. Een subgroep analyse binnen onze cohorte toonde geen verschil tussen de overleving van patiënten met een stentgerelateerde perforatie en deze zonder (p=0.72). Tenslotte was er geen verschil in het ontwikkelen van peritoneale metastasen tussen patiënten met een perforatie en deze zonder perforatie (p=0.47).

DISCUSSIE

Op heden zijn er slechts zeer weinig studies waarin de overleving op lange termijn wordt beschreven na endoscopische stenting van een obstructief coloncarcinoom met heelkunde in tweede tijd. De algemene overleving in onze cohorte is vergelijkbaar met deze die gerapporteerd wordt in de meeste van deze studies. Er is bovendien geen significant verschil te weerhouden wanneer we vergelijken met de algemene overleving van Belgische patiënten met coloncarcinoom, zoals geregistreerd in het Belgisch Kankerregister, hoewel onze patiëntengroep zich initieel presenteerde in obstructie, wat een negatieve prognostische factor is.

De klinische en technische resultaten van onze stenting procedures zijn vergelijkbaar met deze die in de literatuur beschreven zijn. Het relatief hoog aantal perforaties in onze patiëntengroep is te verklaren doordat hieraan bijzondere aandacht werd besteed: niet enkel de klinisch duidelijke perforaties worden gerapporteerd, maar ook de subklinische perforaties die werden weerhouden peroperatief of bij zorgvuldige analyse van het resectiestuk door de anatomopathologen. Er is echter geen significant verschil in overleving te weerhouden tussen patiënten met een perforatie en deze zonder, en ook de incidentie van peritoneale metastasen is niet verschillend. Bovendien is de overleving van deze patiënten met een perforatie gelijkaardig aan de algemene overleving die gerapporteerd wordt door het Belgisch Kankerregister, en dit terwijl deze patiënten in onze cohorte nochtans een slechtere baseline prognose hadden (verder gevorderd tumorstadium en presentatie in obstructie). Dit suggereert dat stentplaatsing op oncologisch gebied veilig is, zelfs indien er een perforatie optreedt.

Bij urgente heelkunde voor een obstructief coloncarcinoom wordt klassiek een stoma aangelegd, dat al dan niet in tweede tijd wordt afgebroken. Het beperkt aantal stoma’s in onze cohorte kan worden verklaard door het electieve karakter van de heelkunde. Verder kan ook een lagere invasiviteitsgraad van heelkunde hieraan worden toegeschreven: de toegangsweg was in de meerderheid van de gevallen laparoscopisch, en slechts bij 3 patiënten werd een subtotale colectomie verricht.

CONCLUSIE

In deze retrospectieve studie rapporteren we (voor zover ons bekend is) de langste follow-up periode die ooit beschreven is in een relatief grote patiëntengroep. Onze resultaten tonen aan dat stentplaatsing voorafgaand aan heelkunde een veilige en effectieve behandelingsoptie kan zijn voor patiënten met een obstructief coloncarcinoom in potentieel curatieve setting.