Next Generation intraoperative Lymph node staging for Stratified colon cancer surgery (GLiSten): a multicentre, multinational feasibility study of fluorescence in predicting lymph node-positive disease

Sentinel lymph node mapping

The ideal solution to the problem of accurate pre-operative LN staging is to develop a reliable means of intraoperative staging, such as sentinel lymph node (SLN) biopsy in breast cancer surgery. This would enable a real-time assessment of LN status. The concept of intraoperative LN staging is not new but has received much interest with the introduction of SLN mapping techniques. The SLN was defined by Morton et al. 19 as the first regional LN encountered by metastasising tumour cells in a study using SLN mapping in melanoma. It was proposed that if the SLN was free of tumour, then it could be assumed that the remaining LNs would also be tumour free and, hence, unnecessarily extensive resections could be avoided in patients with early-stage disease. Early studies of SLN mapping were performed by injecting a dye, such as isosulphan blue or patent blue-V into the primary lesion,19,20 which was taken up into the lymphatics so that the LNs could be visualised at operation. Subsequently, lymphoscintigraphy was added to blue dye, whereby a locally injected radionucleotide was detected intraoperatively using a gamma probe, enabling higher detection rates. 21,22 This concept has proven validity in some cancers, most notably breast cancer, for which the absence of an involved sentinel node spares the patient more radical axillary clearance and associated morbidity. The Association of Breast Surgery Guidelines23 now recommend SLN biopsy for the majority of patients with early invasive breast cancer. SLN mapping was first reported in colorectal cancer in 1999 with blue dye injected around the tumour prior to resection and histological examination of the LNs for cytokeratin expression. The results from early series were not encouraging with SLN detection rates of 58–98%, sensitivity rates of 40–100% and false-negative rates between 0% and 60%. 24 Further studies failed to improve a great deal on these results with large variations in the reported detection rates, sensitivity and false-negative rates, which have been variously attributed to heterogeneity in detection techniques, definition of the SLN, time interval between injection and SLN detection, histopathological techniques, and patient characteristics, including tumour stage and body mass index. 25 A recent meta-analysis showed a SLN identification rate of 92% with a pooled sensitivity rate for detecting LN metastases of 69.6% (range 33.3–100%) and a false-negative rate of 30.4%. 26

Importantly, concerns have been expressed regarding the high false-negative rates and whether metastases from colon cancer follow an orderly spread through tiers of LNs, or rather metastasise as skip lesions. The variability in anatomical site of the first metastatic LN was highlighted by Tan et al. ,27 who found that in 48% of cases the first metastatic LN was not adjacent to the tumour or was 5 cm beyond the longitudinal tumour margin in 18% of cases. Further, Park et al. 28 reported that in 6% of caecal cancers LN metastases occurred along the right branch of the middle colic artery, which might not be included in a standard D2 segmental resection.

Currently, therefore, SLN mapping is not in routine use in colon cancer, although it may have a role in focused histological ultra-staging, whereby positive nodes are subjected to detailed step sectioning in order to increase the detection of micrometastatic disease. 29 The main disadvantage of SLN mapping techniques is that, although they identify the tumour-draining lymphatic basin, they do not discriminate LNs with or without metastatic disease. Therefore, SLN cannot be used for guiding the extent of surgical resection in colon cancer.

Photodynamic diagnosis with 5-aminolevulinic acid

A potential solution to intraoperative LN staging involves drugs used in photodynamic diagnosis and therapy. Probably the most studied drug in this respect is 5-ALA, which is a pro-drug, taken up into the mitochondria of cells, where it serves as a precursor of protoporphyrin IX (PpIX), which, in turn, is the direct precursor of haem. 30 PpIX is a fluorescent molecule which, when exposed to blue–violet light of excitation wavelength 405 nm, emits a characteristic red fluorescence at a wavelength of 630–700 nm.

The 5-ALA is preferentially taken up by cancer cells, which exhibit altered levels of transporter molecules and catalytic enzymes, and metabolised to PpIX. The tendency for cancer cells to accumulate PpIX is enhanced by exogenous administration of 5-ALA. When administered in high doses and irradiated with blue–violet light, 5-ALA is cytotoxic to cancer cells (photodynamic effect), while in lower doses blue–violet light causes the cancers to emit a pink–red fluorescence (photodiagnosis effect).

The reasons for the selective uptake of 5-ALA in tumour cells have not been fully elucidated. In normal cells, 5-ALA synthesis is regulated by a feedback control system that is driven by high intracellular concentrations of free haem. This system is overridden when exogenous 5-ALA is administered. An accumulation of PpIX in a tumour relative to normal cells has been attributed to alterations in the activity of the rate-limiting enzymes porphobilinogen deaminase and ferrochelatase. 30 The former is increased and the latter decreased in tumour cells. Alterations in enzymatic activity may also be due to reduced availability of intracellular Fe²⁺ as a result of rapid cell division. In addition, several biochemical and structural changes in tumour cells have been linked to their avidity for PpIX, including reduced pH, up-regulation of low-density lipoprotein receptors and stromal abnormalities including large interstitial space, leaky vasculature and large amounts of newly synthesised collagens and lipids. 31

Following systemic administration, 5-ALA is metabolised in the liver and excreted in the bile and urine. 30 Much of the 5-ALA is used by the liver to synthesise PpIX and the remaining 5-ALA circulates to other body sites where all other cell types (except those without mitochondria) convert it to PpIX. Intracellular PpIX returns to normal levels up to 48 hours post administration.

A substantial body of work has accumulated on the use of 5-ALA and its derivatives in the fluorescence detection of solid cancers. In humans, 5-ALA fluorescence has been extensively used as a diagnostic aid in transitional cell carcinoma of the bladder,32–35 in neurosurgery to guide malignant glioma resection,36–38 and in gynaecology to detect endometriosis,39,40 peritoneal metastases from ovarian cancer41 and cervical squamous intraepithelial lesions. 42 In urology, several studies have shown the superior detection of transitional cell carcinoma and carcinoma in situ of the bladder using photodynamic diagnosis compared with white-light cystoscopy. 32 Both 5-ALA and its hexyl-derivative, hexaminolaevulinate [(HEX) Hexvix^®, GE Healthcare], have been evaluated in bladder cancer. Burger et al. 33 reported reduced residual tumour and increased recurrence-free survival in patients undergoing photodynamic diagnosis with 5-ALA or HEX than white light alone. Residual tumour was present in 33% of patients with white-light cystoscopy, 15% with 5-ALA and 9% with HEX. The 3-year recurrence-free survival was 67% with white light, 80% with 5-ALA and 82% with HEX and differences between 5-ALA and HEX were non-significant. Similar promising results were reported by Daniltchenko et al. 34 in 115 patients with superficial bladder cancer, but a larger randomised study of 300 patients failed to demonstrate a difference in tumour recurrence. 35 A 2013 systematic review32 showed strong evidence that using 5-ALA or HEX improves tumour detection rates and reduces the residual tumour rate. It also found that, although the majority of the randomised controlled studies demonstrated a reduced local recurrence rate, not all studies confirmed this finding and, therefore, the evidence for recurrence is less convincing.

In a study examining the clinical pharmacokinetics of 5-ALA in normal healthy volunteers and patients at high risk of recurrent bladder cancer, the peak plasma concentration of 5-ALA following oral administration was achieved at 0.83 ± 0.20 hours. Plasma 5-ALA concentrations declined with a terminal half-life of approximately 45 minutes. 43

In neurosurgery, a randomised controlled Phase III multicentre trial36 was performed to investigate the safety and efficacy of fluorescence-guided resection in malignant gliomas. The study involved oral administration of 5-ALA (20 mg/kg) 3 hours (range 2–4 hours) before induction of anaesthesia. A modified neurosurgical microscope was used that allowed switching between conventional white xenon illumination and blue–violet excitation light. Follow-up (median 35.4 months) was available for 139 fluorescence-guided resections and 131 resections carried out with conventional white-light microsurgery. It was shown that 5-ALA fluorescence-guided surgery improved rates of complete resections36 and progression-free survival. 37 It has also been demonstrated in a recent study that there was no difference in the number of adverse events between patients given 5-ALA and the controls. 37 However, a Cochrane review38 of image-guided surgery for brain tumours found that there was low-quality evidence that 5-ALA-guided surgery improves complete tumour resection and the effect on survival is unclear.

In colorectal cancer and its metastases, PpIX has been identified as the predominant endogenous fluorophore and even in the absence of exogenous 5-ALA, it can distinguish involved from uninvolved LNs with a sensitivity and specificity of 62% and 78%, respectively. 44 A total of 33 patients were included in this study, which did not examine intraoperative fluorescence, and the primary tumours and LNs were examined ex vivo. In a study using a murine model of colon cancer, 5-ALA-induced PpIX fluorescence successfully detected LN metastases, whereas benign LNs did not exhibit any apparent fluorescence. 45

In a pilot study involving a mixed group of 18 patients with inoperable adenocarcinomas or adenomas of the oesophagus, duodenum and colon, patients were administered 30–60 mg/kg of 5-ALA orally. 46 In colorectal cancer patients, flexible sigmoidoscopy was performed with serial biopsies. Peak fluorescence in the large bowel tumours was achieved within 6 hours, with good selectivity between tumour and normal mucosa at a ratio of 5 : 1. Similarly effective results were seen in oesophageal and duodenal lesions, but with a lower dose of 30 mg/kg of 5-ALA. In a more recent pilot study, 5-ALA was administered orally at a dose of 20 mg/kg around 4 hours prior to patients undergoing endoscopic resection of early gastric or colorectal tumours. 47 Fluorescence was observed in 6 out of 10 gastric lesions and in one out of the three colorectal lesions. The study noted that fluorescence was observed only in areas of malignant tissue. In this series, four of the patients developed derangement of their liver function tests (LFTs), all of which spontaneously normalised.

A 2013 study examined the ability of 5-ALA to detect LN metastases in freshly excised LNs after en bloc resection of colorectal tumours. 48 The fluorescent images were analysed using a spectral unmixing method to specifically detect PpIX fluorescence. The accuracy, sensitivity and specificity of this method to detect LN metastases were 87.4%, 88.3% and 92.0%, respectively.

An important issue with the clinical use of any photosensitising agent is the potential for side effects. This problem appears to be confined to systemic administration when high doses are used for photodynamic therapy as opposed to lower doses used in fluorescence diagnosis (FD). When used systemically, it is important to establish a safe dose of photosensitiser according to the interval drug to light time to avoid photosensitivity side effects. Normal precautions include keeping the patient out of direct sunlight for 24–48 hours following drug administration. Occasionally reported side effects following systemic administration include nausea, vomiting, tachycardia, hypotension, photosensitivity for up to 48 hours, and elevated liver enzymes. 30 No safety concerns or serious adverse reactions have been highlighted in the literature following the use of 5-ALA in humans.

Primary objective

To optimise the dose of oral 5-ALA administration for intraoperative FD of metastatic LNs in colon cancer.

Secondary objectives

To establish a reliable and repeatable methodology for FD of LN metastasis by standardisation of:

• pre-operative CT LN reporting

• intraoperative fluorescence detection system

• surgical technique for laparoscopic segmental colonic resection with D3 lymphadenectomy

• histopathological examination of resected specimens.

Developmental phase

This would involve a small number of centres to optimise the use and dosage of 5-ALA for intraoperative LN staging in colon cancer and to determine if 5-ALA provided an accurate means for detecting LN metastasis. Two cohorts of 10 patients with positive LNs, as verified on post-operative histology, were treated with different doses of 5-ALA to determine the optimum dose. The end point for this phase was the ability of 5-ALA to detect positive nodes in at least 2 out of the 10 patients with involved LNs in a single cohort. This corresponded to the upper boundary of the 99% exact CI of the sensitivity being at least 60%, that is, a good indication of activity given that the technical protocols and techniques may not have been fully optimised. A 99% CI was chosen to reflect the additional uncertainty in 10 patients. In patients without obvious LN disease at surgery, fluorescence of the primary cancer was used as a positive control for the efficacy of 5-ALA FD in colon cancer. If the initial preliminary phase achieved this end point then the trial would proceed to the evaluation phase.

The most common dose of 5-ALA in the literature is 20 mg/kg and this was chosen as the starting dose given to patients within cohort 1. If 5-ALA produced good fluorescence at this dose, the second cohort was to be treated with a lower dose of 10 mg/kg to reduce the potential of 5-ALA-related side effects. If there was inadequate fluorescence in the first cohort then the dose of 5-ALA was to be increased to 30 mg/kg. The dosing schedule is illustrated below (Figure 1).

FIGURE 1.

The GLiSten study dosing schedule.

By assessing the first 20 patients with positive LNs with this regimen it was hoped that the dose of 5-ALA would be optimised.

For each cohort, particular attention was given to detecting and recording:

• any side effects related to 5-ALA

• intraoperative LN fluorescence

• histopathological confirmation of LN status.

If 5-ALA had detected positive LNs in at least 2 out of 10 patients in either cohort then the optimal dose of 5-ALA would have been identified. At least a further 10 patients with confirmed LN involvement would then be recruited and treated with this optimal dose, with flexibility to include further patients to confirm the validity of the technique before proceeding to the evaluation phase.

Evaluation phase

In this phase, several centres would recruit patients with colon cancer to determine sensitivity, specificity and diagnostic accuracy of 5-ALA intraoperative LN staging compared with in-depth histopathology, creating a much larger cohort.

Participant eligibility

The predefined inclusion and exclusion criteria for the trial are described below.

Potential risks

Participants for this study were selected on the basis that they had a diagnosis of colon cancer that required surgical resection and were suitable for segmental colectomy with D3 lymphadenectomy. There were the normal risks of surgical complications associated with general anaesthesia and surgical resection. There was a small risk of photosensitivity reactions related to 5-ALA administration. This included transient derangement of LFTs (expected to return to baseline at 24–72 hours post administration) and skin hypersensitivity reactions owing to exposure to ultraviolet light. Other mild and transient side effects associated with 5-ALA administration in the literature include nausea, vomiting, tachycardia and hypotension. When 5-ALA is administered in conjunction with surgical resection of malignant brain tumours, a potential side effect is brain oedema and, although this was felt to be an unlikely occurrence in our trial patients, this information was included in the trial patient information leaflet.

In response to these potential adverse effects, all patients were kept out of direct sunlight for at least the first 48 hours following surgery and monitored for changes in vital signs and liver function (which is part of the standard post-operative care). During the operation, patient’s eyes and skin were protected from the operating lights using standard methods such as sterile drapes and tape to keep their eyes closed.

Concurrent clinical trials

Patients were screened for participation in other clinical trials. Providing there was no conflict, then patients could be included in both GLiSten and other trials. Notable exceptions were patients recruited to FOxTROT in which pre-operative chemotherapy is given. This can downstage the disease and, therefore, may have reduced the probability of detecting LN metastases.

Duration of participant participation

Patients were followed up for the first 30 post-operative days and this completed their participation in the trial. After this point, patients continued with standard follow-up for their colon cancer as per their institution’s normal clinical practice.

Withdrawals or removal of patient criteria

The right of the patient to refuse consent without giving reason was respected and all patients were free to withdraw from the study at any time without giving reasons and without prejudicing any further treatment.

Pre-operative care

All participants underwent a pre-operative staging CT of the thorax, abdomen and pelvis as standard practice within 8 weeks of surgery. All other pre-operative assessment was as per routine cancer practice including colorimetric tattooing of the tumour at colonoscopy. Oral 5-ALA was prepared as described in Appendix 3 and administered to the patient 1–6 hours prior to surgery.

Perioperative care

Participants were initially assessed with blue-light laparoscopy to detect any fluorescence within the tumour and draining lymphatic field within the mesentery. Fluorescent LNs were marked with Ligaclips [Ethicon Endo-Surgery (Europe) GmbH; Norderstedt, Germany] to facilitate subsequent pathological identification. Patients then underwent a segmental colectomy with D3 lymphadenectomy, if appropriate. For cancers of the right colon, segmental colectomy with D3 lymphadenectomy aimed to include central dissection at the origin of the ileocolic, right colic (when present) and middle colic vessels, with high vascular ligation and division. For cancers of the sigmoid colon, segmental colectomy with D3 lymphadenectomy included high ligation and division of the inferior mesenteric vessels proximal to the origin of the left colic vessels. Specimen extraction, anastomosis and formation of stoma were left to the discretion of the operating surgeon.

The sites of any fluorescent nodes were marked at operation with small surgical clips to facilitate the pathologists’ assessment. Resection specimens were subjected to in-depth histopathology, as per the Royal College of Pathologists guidance49 as well as additional assessment of fluorescent LNs in the form of step sectioning. This included:

• specimen photography and assessment of the completeness of mesocolic resection of the fresh and the fixed specimen

• distance of tumour and bowel wall to vascular ligation and length of vessels

• area of lymphadenectomy

• level of lymphadenectomy based on vascular anatomy

• histological analysis and mapping of all LNs.

Post-operative care

Participants received standard post-operative care with additional monitoring for side effects to 5-ALA, including measurement of liver and kidney function tests. Patients were reviewed at 30 days postoperatively, as per standard clinical practice. This marked the end of their participation within the trial but they continued to receive appropriate follow-up care as per guidelines. 50

Investigational medicinal product supply, labelling, handling and investigational medicinal product preparation

5-aminolevulinic acid was supplied by photonamic GmbH & Co. KG with a trial-specific label in line with Directive 2001/20/EC and the Medicines for Human Use (Clinical Trials) Regulations 2004 (amended 2006). 5-ALA was prepared and handled in line with manufacturers’ recommendations.

5-aminolevulinic acid administration

5-aminolevulinic acid was supplied in a sterile powder format within individual vials. Each vial contained 1.5 g of 5-ALA and was reconstituted in the vial by adding 50 ml of tap water using a needle and syringe, giving a concentration of 30 mg/ml. This was done immediately prior to oral administration.

The patient’s weight was used to calculate the amount of 5-ALA to be given, which was checked and verified by the hospital pharmacy department prior to them dispensing the drug the day prior to surgery. On the morning of surgery, the drug was reconstituted as above by the clinical research fellow and witnessed by nursing staff who countersigned the prescription. The correct volume of the reconstituted 5-ALA was calculated and then given to the patient.

Concomitant medications

The only absolute contraindications to the use of 5-ALA are:

• a previous history of hypersensitivity to photosensitising drugs

• a previous history of acute or chronic porphyria

• hepatic or renal dysfunction

• pregnancy.

5-aminolevulinic acid should be used with caution:

• with other medications known to have a photosensitising effect, for example tetracylines, sulphonamides, quinolones. Patients should not be exposed to any photosensitising agents up to 2 weeks after administration of 5 ALA

• with other medications associated with acute porphyria, for example diclofenac, barbiturates, carbamazepine, phenytoin

• with other medication associated with hepatic or renal dysfunction, for example non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors, loop diuretics, phenytoin.

Cohort 1

In the first cohort, 18 patients were treated with 20 mg/kg of 5-ALA but only 17 were operated on with blue-light laparoscopy, as one patient’s operation was cancelled. Six out of the 17 patients had a fluorescent primary tumour (35.3%) and three of these patients had fluorescent LNs (see Table 2).

One out of the three patients with fluorescent LNs had confirmed metastatic nodal disease within the fluorescent node on standard histopathology. Therefore, there was only one confirmed case of a fluorescent positive node out of 10 patients with positive LN disease. The success detection criteria (the ability of 5-ALA to detect positive LNs in at least 2 out of 10 node-positive patients) for cohort 1 was therefore not met.

In cohort 1, 5-ALA has a sensitivity of 11.1% and specificity of 75% for determining the pathological outcome of LN assessment using fluorescence (Table 3).

Two other patients in cohort 1 had fluorescent nodes; however, none of these fluorescent nodes contain any metastatic disease on standard histopathological examination. One patient had no detectable metastatic disease in any of their LNs including the fluorescent node on standard histopathological examination. Therefore, this patient is counted within the fluorescent node-negative group. The third patient had metastatic disease present in 6 out of 23 nodes found within the pathological specimen; however, no metastatic disease was found within the fluorescent node on standard histopathological examination. This patient could occupy two groups in Table 3: the fluorescent node-negative group or the non-fluorescent node-positive group. As we are primarily interested in the outcome of the fluorescent nodes, this patient was counted in the fluorescent node-negative group as they are considered to be false positive for the purposes of statistical analysis. However, this does mean that the total number of node-positive patients in Table 3 will appear to be nine instead of 10.

In summary, in cohort 1, 18 patients were treated with a dose of 20 mg/kg of 5-ALA and only one-third of patients had a fluorescent primary tumour. A total of 10 out of the 18 patients were found to be node positive. Only 1 out of these 10 patients had correctly identified node-positive disease via fluorescence, that is, a positive diagnosis. Therefore, cohort 1 did not meet the predefined criteria for acceptable 5-ALA sensitivity (2 out of 10) and the dosage was increased to 30 mg/kg, as per the protocol.

Cohort 2

Twenty-six patients were recruited to cohort 2 and nine of these patients had confirmed nodal disease on standard histopathology. Only 24 patients underwent blue-light laparoscopy because one patient had their operation cancelled and the equipment was not available on the day of surgery for another. A third patient was found to have unresectable disease but blue-light laparoscopy was performed. The tumour exhibited fluorescence, however none of the LNs fluoresced in this patient; therefore, only 23 patientsunderwent blue-light laparoscopy and had histopathological examination of the tumour performed.

Eight out of the 24 patients who underwent blue-light laparoscopy had a primary fluorescent tumour (33.3%) and four of these cases also had fluorescent LNs. None of the patients with fluorescent LNs was found to contain metastatic disease on standard histopathological examination (Table 4).

In cohort 2, 5-ALA has a sensitivity of 0% and specificity of 75% for determining the pathological outcome of LN assessment using fluorescence (see Table 4).

Out of the four patients in cohort 2 with fluorescent LNs, two had no evidence of metastatic disease in any of the LNs within the specimens and are, therefore, counted in the fluorescent node-negative group. The other two patients had non-fluorescent nodes containing metastatic disease but the fluorescent nodes were benign on standard histopathological examination; therefore, similarly to the patient in cohort 1, they have been counted in the fluorescent node-negative group instead of the non-fluorescent node-positive group. This means that the total number of patients with node-positive disease in cohort 2 will appear to be seven instead of nine, according to Table 4.

In summary, in cohort 2, 26 patients were treated with a dose of 30 mg/kg of 5-ALA. Nine out of the 26 patients were found to be node positive. None of the nine patients had correctly identified node-positive disease via fluorescence, that is, a positive diagnosis. Recruitment to the trial ceased prior to reaching the target of 10 patients with node-positive disease in cohort 2; however, it is clear that the predefined criteria of at least two correct diagnoses in the 10 node-positive patients could not be reached as there had not been one correct diagnosis in nine node-positive patients.

Additional histopathological studies of all fluorescent benign LNs in both cohorts are continuing. Further staining of the LNs has not shown any evidence of metastatic disease but the results of immunohistochemistry are still awaited.

Summary

In conclusion, 7 out of 19 patients with node-positive disease exhibited LN fluorescence (36.8%). Neither dose of 5-ALA had adequate sensitivity to meet the defined end point and allow progression into the developmental phase of the trial.

Quality of pathology specimens

The oncological quality of the resected specimen was judged on the plane of resection, which was reported in 40 pathology specimens. A total of 26 out of the 40 specimens were resected in the mesocolic plane (65%). There was a higher proportion of mesocolic resections in cohort 2 than cohort 1: 78.2% versus 47.1%, respectively (Table 8). These figures should be interpreted in light of the enriched cohort of locally advanced colon cancers. Although the 65% for complete mesocolic excision is lower than reported elsewhere in the literature, the advanced nature of the cancers in this study will have directly impacted on the ability to perform complete mesocolic plane surgery.

Resection margins

A total of 10 cases have been reported as having positive microscopic resection margins (R1), resulting in a positive resection margin rate of 23.8%, which is below the rate deemed to be excessive in the protocol (> 30%). Although this is higher than the R1 resection rate normally reported in colon cancer surgery, it probably reflects the enrichment of the cohorts with advanced stage disease and that some pathologists classify T4 disease with serosal involvement as a positive resection margin and would not be a cause for concern. There were no R2 resections in either cohort.

Complications

The overall anastomotic leak rate was 4.5%, which is below the rate deemed to be excessive in the trial protocol, but in keeping with the literature for this type of surgery. In the first cohort, four laparoscopic cases were converted to open laparotomies, resulting in a conversion rate of 22%, which is below the rate deemed to be excessive in the protocol (> 50%). There were two conversions in the second cohort giving a conversion rate of 8.3%. Although a conversion rate of 22% in cohort 1 would ordinarily be viewed as high, this cohort was enriched with advanced cancers and within this context a 22% conversion rate is probably acceptable.

There have been three intraoperative complications: one visceral perforation and two cases of intraoperative haemorrhage. In the first case, there was a perforation to the greater curve of the stomach following introduction of the Veress needle – this was repaired laparoscopically and the patient made an uneventful recovery. In the first case of intraoperative haemorrhage, there was bleeding from the inferior vena cava, which was clipped laparoscopically and no conversion was necessary; the patient made an uneventful recovery. In the second case of intraoperative haemorrhage, blood was seen in the abdominal cavity during closure of the laparoscopic port sites. The surgeon converted to an open procedure and found that the source of bleeding was from a peripancreatic vessel. This was oversewn and haemostasis achieved. The patient was admitted to the high-dependency unit (HDU) postoperatively for observation and made a good recovery. All three of these intraoperative complications are recognised within the type of surgery performed.

Two other patients were managed for post-operative ileus in addition to a patient who had to be readmitted, all of whom had CT scans during the initial few days of their post-operative course. One of these patients was treated for a chest infection with intravenous antibiotics. The other patient with a post-operative ileus underwent CT and the scan did not show any pathology. However, the patient became seriously ill with septic shock in the intensive care unit (ICU). A decision was made to undertake exploratory laparotomy, despite the normal CT scan, and the patient was found to have a sigmoid volvulus causing ischaemic perforation of the bowel. A completion colectomy with end ileostomy was performed. The patient subsequently had a prolonged stay in the ICU, during which he sustained a left internal jugular vein thrombosis secondary to central line placement requiring treatment with low-molecular-weight heparin.

Thirty-day mortality

Unfortunately, one patient died within the first 30 days of surgery. This patient had significant comorbidities and was identified as a high-risk surgical candidate on pre-operative cardiopulmonary exercise testing. The risks were discussed with the patient and their family and the patient was keen to proceed with surgery, including participation within the GLiSten study. The patient developed a lower respiratory tract infection in the early post-operative period but responded well to conservative management. Unfortunately, the patient was found to be acutely peritonitic > 10 days post surgery and findings at relaparotomy were consistent with an anastomotic leak. The anastomosis was resected and a defunctioning stoma was created. Subsequently, the patient suffered a non-ST elevation myocardial infarction and went on to develop multiorgan failure, to which he succumbed. This case was reviewed by the Data Monitoring and Ethics Committee (DMEC) and TSC and 5-ALA was not felt to have been implicated in the patient’s development of complications and demise.

Readmissions

Three patients had to be readmitted to hospital shortly following their discharge from the surgical ward. The first patient developed a lower respiratory tract infection that required treatment with intravenous antibiotics. This is a well-recognised post-operative complication and was not thought to be related to 5-ALA. The patient responded well to antibiotic treatment. The second patient requiring readmission presented with abdominal pain and acute kidney injury. The patient had urea and electrolyte levels checked for the first 5 post-operative days and, apart from a slight increase in creatinine on day 4 which had normalised by day 5, no other abnormalities were detected. The kidney injury was believed to be secondary to concomitant medications, which can cause renal impairment, and not as a consequence of receiving 5-ALA. The patient’s renal function recovered with intravenous fluids and no other intervention was required. The third patient was managed for a post-operative ileus and a CT scan showed a small amount of intraperitoneal gas, with no demonstrable intra-abdominal collection. It is unclear whether or not this patient had a small anastomotic leak, but the patient was managed conservatively with oral antibiotics for 7 days and was readmitted to hospital for 2 days following their initial discharge.

Drug safety

There have been no drug-related serious adverse events (SAEs) but we did see some minor 5-ALA side effects in eight patients. There were two patients, both male in cohort 2, who developed mild self-limiting photosensitivity reactions that affected the head and neck region. Both of these patients were admitted to the HDU postoperatively for observation and it must be noted that HDU has a much brighter ambient environment than the normal surgical ward. In both cases, the reaction was noted within the first 48 hours following surgery and had resolved by the time of discharge (around post-operative day 5). It is proposed that the brighter ambient environment in the HDU is likely to be a contributing factor. No intervention was required in either patient other than a topical emollient cream for associated skin dryness.

Four patients were noted to have elevated LFTs in the immediate post-operative period. No intervention was required, apart from monitoring, and the LFTs normalised within 3 days post surgery. Two patients had a slight transient deterioration in their renal function but it is not clear whether this was related to the surgery or the 5-ALA. Either way, renal function returned to normal in both patients and without any intervention.