Follicular lesions of the thyroid: a surgical perspective.

Editor,

The management of follicular thyroid lesions has long been a challenge for the surgeon due to the difficulty of differentiating pre-operatively between adenoma and carcinoma. Even with advances in ultrasonography, fine needle aspirate cytology, (FNAC), and immunohistochemistry we are still relying on post-operative pathological analysis for a definitive diagnosis. To date no single test or combination can accurately predict the presence or absence of vascular or capsular invasion in these lesions.1

The aim of this study was to investigate the accuracy of FNAC and its influence on the surgical management in this patient population. We collected data for confirmed follicular adenoma or carcinoma on tissue diagnosis over a 5 year period, 2008-2012, from the pathology laboratory of the Belfast City Hospital. We retrospectively analysed the pre-operative FNAC for predictive diagnostic accuracy.

From this 5 year data pool there were a total of 53 follicular thyroid lesions identified on final histology, (45 adenoma and 8 carcinoma). Of these 5 FNAC reports were not available or non-diagnostic, (3 adenoma and 2 carcinoma), giving a total of 48 complete data sets. The FNAC were divided into two categories, benign (Thy 2) and suspicious of malignancy, with all results that couldn’t safely be considered benign being consigned to the latter (Thy 3,4&5). It should be noted that all lesions identified as being follicular in nature should be recorded as Thy3.

Data from Thy0 and Thy1 were excluded from subsequent calculations due to lack of FNAC direction. Anything that could not be definitively said to be benign was grouped as suspicious of malignancy as from a surgical perspective all of these cases would be offered excision.

From the results we can see that the inclusion or exclusion of follicular lesions (Thy3 category) vastly alters the specificity, positive predictive value and overall diagnostic accuracy of the data. This is particularly relevant to the surgical workload as anyone with a follicular lesion will be offered, as a minimum, a thyroid lobectomy. If this proves to be malignant on histology, following discussion at the Multi-Disciplinary Meeting, some patients will then be offered a completion total thyroidectomy.2 This increased workload is further compounded by the increasing number of asymptomatic or subclinical thyroid incidentalomas being diagnosed when imaging is performed for other reasons. The incidence of these lesions is high, with various ultrasound studies suggesting 10-67%.3

With growing waiting lists, the current economic climate and operating time being a restrictive commodity, better investigations need to be found to reduce the amount of unnecessary surgery being performed. Whilst this is true for all thyroid incidentalomas it is particularly so when it comes to follicular thyroid lesions.

Editor,

Invasive coronary angiography is the current gold standard in the work-up of non-specific chest pain suggestive of ischaemic heart disease.1 This case demonstrates how CT coronary angiography can yield a clinically useful answer when traditional catheter angiography runs into difficulty.

A 61 year old female smoker with hyperlipidaemia presented with a four month history of anginal chest pain and dyspnoea with a positive cardiac stress ECG. This was investigated by invasive coronary angiography, this proved to be challenging due to failure to cannulate the left main coronary artery. However during angiography a large right coronary artery was noted with multiple collateral vessels perfusing the left anterior descending coronary artery territory.

Further anatomical detail of the left coronary artery was revealed through coronary CT angiography which showed a diffusely occluded left main stem with no definite communication with the aorta. Perfusion was maintained by long standing collateralization by vessels originating from a large right coronary artery. The patient went on to receive coronary artery bypass grafting of her left coronary system via a left internal mammary artery graft to the LAD.

In this case, coronary angiography failed to accurately identify pathology in this case due to an inaccessible left coronary artery. As highlighted by figure 1 the left coronary artery was partially visible through contrast supplied from collaterals originating from the right coronary artery. However a significant coronary stenosis could not be excluded. Moreover this patient was not amenable to percutaneous coronary interventions in view of the difficult angiography.

Diagnostic Images

1A: Invasive Coronary Angiography. Large Pointer - Conus branch collaterals to left coronary sinus. Small Pointer - Left anterior descending artery territory

1B: Coronary CT angiography axial MPR Pointer – Remnant Left main stem

1C: Coronary CT Angiography 3D reconstruction Small Pointer: Collaterals from the RCA perfusing LAD Large Pointer: Left anterior descending artery

A coronary CT angiogram was performed in order to exclude coronary artery disease and provide anatomical detail for possible surgical intervention. As shown in the figure this technique revealed a completely occluded left main stem. There was a very small caliber left main coronary and proximal left anterior descending artery. The caliber of the LAD then became normal with multiple collateral vessels from the right coronary artery providing collateral antegrade and retrograde flow. From this information obtained, the patient further underwent a left internal mammary bypass graft to her left circumflex artery and was subsequently asymptomatic.

This case report highlights the utility of multi-modal cardiac imaging in cases with unusual coronary anatomy and inconclusive invasive angiography and thus aids in clinical decision making.1

Editor,

The management of thoraco-abdominal aneurysms (TAA) remains a surgical challenge with high rates of morbidity and mortality.

Spinal cord ischaemia with subsequent neurological sequelae is a recognised complication of these repairs, with most of this morbidity occuring acutely. There is no completely satisfactory method of protecting the cord during repair. One strategy is to pre-operatively visualise the segmental artery supplying the Adamkiewicz artery and ensure its revascularisation. Alternatively, a catheter may be placed into the intrathecal space to assess CSF pressure, allowing free drainage of CSF to a pressure below 10mmHg1.

We report delayed cord ischaemia in a 76-year-old man with a previous open abdominal aortic aneurysm (AAA) repair, presenting with a Crawford Type II TAA.

A hybrid endovascular and open repair was undertaken. Prior to induction, a spinal catheter was inserted, allowing free CSF drainage to below a pressure of 10mmHg. Open visceral revascularisation was performed using a four-limbed graft from the existing distal infra-renal aortic graft to the left and right renal arteries, superior mesenteric artery and coeliac axis (Figure 1). Endovascular TAA stent grafting was then performed via an open aortic graft conduit using a Relay® thoracic stent. The graft extended proximally beyond the left subclavian and distally to the infra-renal aorta. The APTT was maintained at 2-3 times normal range throughout with heparin.

Once extubated, lower limb motor and sensory function were confirmed as normal and there was no evidence of organ dysfunction.

Twenty-four hours post-operatively the patient developed bilateral lower limb flaccid paralysis, with sensory and motor deficit from T8 downwards. Despite spinal fluid drainage via the spinal drain, there was no improvement. CT showed no evidence of epidural haematoma but showed spinal cord infarction and oedema distal to T8. MRI confirmed anterior pattern spinal cord infarction distal to T8. The patient recovered from surgery but had persistent paralysis of the lower limbs, unchanged at the 12 month review.

Type II aneurysms have a significantly greater risk of paraplegia, with up to 22% of patients presenting with paraplegia up to 3 months after seemingly successful surgery1. A series of 89 urgent and elective high-risk patients in Europe demonstrated an 8% risk of paraplegia1. Crawford et al., observed an immediate neurological deficit in 68% and a delayed deficit in 32% of patients affected2. Murphy et al., found no difference in spinal cord ischaemic rates between patients undergoing hybrid repair and those undergoing open repair3.

Adjunctive techniques are used to reduce the risk of cord ischaemia. A randomised controlled trial by Coselli et al., demonstrated an 80% reduction in the relative risk of postoperative paraplegia when the technique of CSF drainage was implemented4. Safi et al., showed an acute CSF pressure increase in patients prior to development of spinal cord paralysis, preceeded by a period of blood pressure instability5.

We suggest that delayed spinal cord ischaemia can still occur in a hybrid type repair despite adjunctive CSF drainage. Morbidity and mortality rates are improving but surgery for TAA still carries significant risk.

Editor,

Following the recent focus on campylobacter during Food Safety Week 20141, it would be timely to inform readers of recent changes in Public Health investigation of campylobacter cases.

Campylobacter remains the most commonly reported cause of food poisoning. The main sources include raw or undercooked meat (especially poultry), unpasteurised milk and untreated water2.

Presently, campylobacter infections are reported by microbiology laboratories and clinical teams to the Health Protection Service at Public Health Agency (PHA). Campylobacter reports in Northern Ireland have increased from 843 cases in 2008 to 1350 cases in 2013. Similar increases in campylobacter cases have been seen in all areas across UK. This may be due to increasing sensitivity of testing methods used by laboratories, however, ultimately the cause is unknown.

Until recently all campylobacter cases reported to PHA were sent a postal questionnaire. Completed questionnaires were then reviewed to identify infection control measures required and to identify outbreaks promptly.

An audit was undertaken to review current practice and identify potential service improvements in this area of health protection.

The audit looked at Campylobacter cases over a 6 month period in 2013. Over this time period there were a total of 837 laboratory notifications with 403 (48%) cases returning questionnaires.

Median time from onset of symptoms until confirmation of infection was 9 days and median time between laboratory confirmation and receipt of completed questionnaires was also a further 9 days.

Majority of questionnaires (96%) did not identify any risk factor exposures, further cases or outbreaks. Of the remaining cases, only 1% led to environmental inspection and sampling of food premises.

On further discussion with other health protection units across the UK, it became clear that there is a wide variation in approach to investigation of campylobacter. Currently there is no evidence to identify the optimal approach to investigation of this disease.

Several limitations were identified from our audit. These included the immediate loss of information on 52% of cases and the significant time delay of 18 days between onset of symptoms and review of case information. Given campylobacters’ relatively short incubation period of 3 days on average, these responses were not timely enough to facilitate infection prevention or early identification of outbreaks.

The fact that only 4% of cases required further action revealed an obvious mismatch between the investigative resources required and Public Health action taken.

Based on this audit a new approach to public health investigation of campylobacter was instigated in July 2014, adapted from Good Practice statements published elsewhere3.

Cases are now reviewed and classified as Not linked, Possibly linked or Probably linked dependant on a number of factors, with a graded response for each. This simple change in clinical practice has led to service improvements from both temporal and financial perspectives.

The PHA would continue to encourage all health professionals in primary and secondary care to be aware of food-borne pathogens. If enteric infection is suspected we would recommend that microbiological testing is requested and that all symptomatic patients are given general infection prevention control advice.

The authors have no conflict of interest.

Editor,

Ileal conduits are used for urinary diversion following cystectomy for trauma, malignancy, congenital defect or neurogenic non-functioning bladder. Urinary diversion tumours are usually in uretero-sigmoidostomies where mixed faeces and urine may be tumourogenic.1 Tumours in transpositioned small bowel segments are usually found post-cystectomy seeded from the primary2. Primary malignancy of the small bowel is uncommon, with 175 cases in the UK over two years. Primary tumours in an ileal conduit are exceptionally rare2.

We present two patients with primary adenocarcinoma in their ileal conduits.

This 78-year-old woman underwent radical cystectomy and ileal conduit for transitional cell carcinoma (T2N0, Grade 3, CIS bladder only). She had lifelong ciclosporin post cardiac transplant for cardiomyopathy. On admission for bleeding she developed abdominal pain, vomiting and non-functioning stoma. Examination revealed abdominal tenderness and a parastomal hernia. CT scan demonstrated thickened omentum, parastomal hernia, ileal conduit dilatation and small bowel obstruction.

At laparotomy, an ileal conduit tumour was found, adherent to transverse colon, with widespread tumour deposits. The ileal conduit was taken down and refashioned and right hemicolectomy and ileostomy was performed.

Pathology confirmed poorly differentiated adenocarcinoma with desmoplastic stroma arising from the conduit wall involving the ileum, colon and omentum (Figure 1). The patient died from cardiogenic shock within 24 hours.

At  laparotomy,  an  ileal  conduit  tumour  was  found,  adherent  to  transverse  colon,   with  widespread  tumour  deposits.    The  ileal  conduit  was  taken  down  and   refashioned  and  right  hemicolectomy  and  ileostomy  was  performed.

Pathology  confirmed  poorly  differentiated  adenocarcinoma  with  desmoplastic   stroma  arising  from  the  conduit  wall  involving  the  ileum,  colon  and  omentum   (Figure  1).  The  patient  died  from  cardiogenic  shock  within  24  hours.

A low and high power photomicrograph of poorly differentiated adenocarcinoma arising from the wall of the conduit.

A 45-year-old woman with congenital ectopia vesica had cystectomy and ileal conduit formation at age five. Later, she had multiple gynaecological procedures, strangulated hernia repair, division of adhesions, hysterectomy and excision of the resultant enterocutaneous fistula.

Presenting with abdominal pain, CT scan showed an obstructing lesion in her urostomy and hydronephrosis. At laparotomy, the conduit was excised and a new one fashioned. Pathology demonstrated a mixed neuroendocrine tumour and mucinous adenocarcinoma. She had adjuvant chemotherapy and remains disease free.

Primary adenocarcinoma in an ileal conduit is rare with the absolute risk unknown2. Carcinogenic nitrosamines, increased oxidative stress or release of inflammatory mediators (epidermal growth factor, cytokines and cyclo-oxygenase-2) as a result of chronic inflammation or recurrent infections are possible mechanisms of carcinogenesis3. Biopsies of ileal mucosa in patients for 7 years post-surgery found mucosal thinning and villous atrophy but no malignancy.3 This may have underestimated the latent-period which can be up to 40 years.

Small bowel adenocarcinoma is treated primarily by surgery, however, many present late.4 Limited data suggests improved survival with adjuvant chemotherapy.4

There is no evidence that immunosuppression increases the risk of malignancy in ileal conduits. However, an increased risk of malignancy in renal transplant and rheumatoid arthritis patients, treated with azathioprine, has been observed, with case reports of cancers in patients with Crohn’s who receive immunosuppression.5

These cases highlight the potential for malignancy in ileal conduits. A high level of suspicion is appropriate in patients with urinary obstruction or pain in the urostomy. They require radiological and endoscopic evaluation for recurrence or primary malignancies of the conduit. Patients with longstanding urinary diversion may warrant surveillance, however there is no consensus regarding this.

Editor,

Small bowel malrotation (SBM) is a congenital anatomic anomaly resulting from an abnormal rotation of the midgut during embryogenesis which places the patient at risk of acute and chronic complications. SBM is historically regarded as a paediatric phenomenon, however over the past 30 years literature has increasingly focused on the incidence, significance and management of SBM within the adult population. Traditionally prevalence is often quoted as ∼0.2% however authors of more recent studies agree the true incidence of malrotation in adults is underestimated due to the wide range of potential clinical presentations1.

Symptoms are often mistaken for irritable bowel syndrome, peptic ulcer disease, biliary and pancreatic disease and psychiatric disorders. A number of studies have noted those initially classified as ‘asymptomatic’ have in fact on careful questioning reported abdominal complaints attributable to their malrotation1. Therefore an abnormal junction in an adult should not be dismissed simply as a normal variant.

Upper GI contrast fluoroscopy remains the examination of choice, however with increasing utilisation of CT more recent studies have observed features such as an abnormal SMA/SMV relationship or associated extra-intestinal abnormalities such as aplasia of the uncinate process of pancreas2.

The greatest controversy surrounds the management of the asymptomatic patient who is discovered incidentally to have malrotation on radiologic examination. The literature is divided with some authors advocating a surgical Ladd procedure only in symptomatic patients. In 2006 Malek and Burd3 undertook a unique ‘decision analysis’ study to assess the risk of surgery compared with “watchful waiting.” They concluded observation of completely asymptomatic malrotation in adults older than 20 years of age resulted in increased life expectancy in those individuals.

In contrast, other authors have argued any risk of midgut volvulus, regardless of how small, warrants operative intervention4. In the absence of an adequately sensitive or specific tool to predict those at risk of developing such complications, some authors argue surgical intervention should be considered. Authors have also suggested increased access to laparoscopic Ladds procedures, which are associated with reduced post-operative morbidity, should encourage prophylactic correction of asymptomatic SBM5.

An audit of local practice revealed adequate visualisation of the duodeno-jejunal junction on 92% of audited barium swallows and meals. However only 14% of adult patients with a radiologically diagnosed SBM were referred to a surgical team. As a result of this audit it has been agreed a departmental standardised protocol for upper gastrointestinal barium studies should be created and the local upper gastrointestinal surgeons have supported the referral of adult patients with radiologically diagnosed SBM.

The authors have no conflict of interest.

Editor,

The laparoscopic approach to gallbladder surgery has almost completely replaced conventional open cholecystectomy (OC) as the gold standard for symptomatic cholelithiasis. The open approach is generally reserved for complex cases with unclear anatomy or intra-operative complications that cannot be managed laparoscopically1, 2.

The combination of low conversion rates to OC and decreased training hours due to the introduction of the European working Time Directive has led to a hypothesis that current trainees may not be competent to perform open cholecystectomy1-3.

As the need for conversion to OC is not always predictable, and since experienced colleagues may not be at hand, this is a potentially important clinical governance issue.

We set out to evaluate the experience and competencies of higher surgical trainees in Northern Ireland in the performance of OC.

An email survey of all higher surgical trainees in Northern Ireland was conducted. Each trainee provided their year of training, along with details of how many LC’s and OC’s they had assisted with, performed with assistance or performed independently. Trainees’ confidence to perform the 12 different steps of an OC, as outlined on Intercollegiate Surgical Curriculum Project (ISCP) assessment form were also assessed. Trainee opinion on the available options to enhance OC training was also sought.

Twenty seven (79%) of the higher surgical trainees responded to the survey. The level of training covered the full spectrum of Specialist Training (ST) 3 to ST 8, with 11 (41%) responders within their final 2 years of training.

The survey data is summarised in Table 1.

8 out of the 10 trainees who had performed more than 10 OC’s under supervision were in the last two years of their training. Of these 8 trainees, 5 had rotated through an HPB unit.

When asked about their confidence in performing all steps of an OC, 7 trainees were confident of carrying out all steps independently. The most junior trainee confident at performing an OC independently was ST6 level. 9 trainees did not feel confident to perform all steps even with assistance. Of the 9 trainees who had passed through the regional HPB unit, 6 were confident of all steps of OC.

18 (67%) trainees felt they should not be performing LC independently if they were not confident to perform OC. 16 of these 18 trainees plan to perform LC’s independently in the future

All trainees felt that through attending theatre lists and rotating through HPB, their ability to perform OC would improve. Simulator practice and watching videos were identified as potentially useful adjuncts to open cholecystectomy training.

The current open surgical simulation studies show, in general, a benefit in developing the surgical skills of surgical trainees. However, these studies do have limitations. There are no current simulators in the literature for training in open cholecystectomy. In hepatobiliary surgery, simulators do exist for t-tube insertion and are being developed for practising forming a choledochojejunostomy and pancreaticojejunostomy4, 5. We feel there would be a benefit in developing a simulator for the teaching and practice of open cholecystectomy.

In conclusion, whilst it is of paramount importance to competently perform a laparoscopic cholecystectomy by the completion of higher surgical training, the core skills of performing an open cholecystectomy in a complex situation need to be maintained. Training therefore needs to address the needs of the trainees through rotation through an HPB unit, and simulation.

Data, excluding Thy3 category	All data
Sensitivity = 100%	Sensitivity = 100%
Specificity = 75%	Specificity = 22%
Positive predictive value = 40%	Positive predictive value = 18%
Negative predictive value = 100%	Negative predictive value = 100%

Table 1:

Number of procedures performed by surgical trainees

Number of trainees who performed each procedure	Number of Procedures
	0-10	11-20	21-30	31-40	≥40
LC’s assisted	-	-	2	3	22
LC’s performed under supervision	1	1	1	2	22
LC’s performed independently	8	6	2	3	8
OC’s assisted	5	12	5	4	1
OC’s performed under supervision	17	8	1	1	-
OC’s performed independently	27	-	-	-	-