'High risk medicolegal autopsies: is a full post-mortem examination necessary?'

Comment on the article by Fryer et al, J Clin Pathol 2013, 66:1-7

The article by Fryer et al raises several critical issues - I do not agree with them on all points - and leads to an important overall conclusion for the future prosecution of autopsies in the UK.

The piecemeal introduction of cadaveric imaging for non-homicide cases in (currently) a few centres in England is probably unstoppable, and has the backing of government (although not the central funding). As stated in their article, it is intended to avoid performing an open autopsy examination in a proportion of cases where a coroner has commanded a post-mortem examination. Interestingly whilst this article, and other publications like it, discuss replacing open autopsy with imaging, in public educational fora when presentations by pathologists and radiologists are made, the emphasis is more on imaging as an adjunct than a replacement for autopsy. What is not widely discussed anywhere is how the useful contribution of cadaveric imaging is critically dependent on clinical case-mix.

We would all agree that having imaging data prior to commencing a standard autopsy, whether the imaging was done under hospital care prior to death or done just prior to the autopsy as with deaths in the community, is valuable. It focuses attention to relevant clinical pathologies to be examined, and provides negative information for some organs (such as the brain in ?intracranial haemorrhage). An unintended and underused consequence of pre-autopsy imaging is enabling feedback to radiologists, particularly in the in-hospital setting, of their diagnostic performance. They necessarily receive plentiful such information in cancer multi-disciplinary meetings, but do not regularly obtain information on their diagnostic hits and misses when it comes to the moribund. Like all diagnosticians they are fallible.

High risk cases Fryer et al suggest that in cadavers of patients with suspected illicit drug intake, and known 'high-risk infection' (specifically Hazard Group category 3 infections, including HCV, HBV, HIV), an external examination plus blood sample toxicology can remove the need for autopsy in more than half the cases. The results, presented in % terms, are indeed persuasive. Toxicology alone provided the cause of death in 78% of cases, and CT scan alone in 25%. In the validation group (suspected drug abuse but no infection), the toxicology provided the cause of death in 87%.

The arguments presented for avoiding open autopsy on high risk cases include expense, disruption to busy mortuary work, and health risk to staff. Interestingly, the authors state that some pathologists in their centre are reluctant to perform autopsies on such cases. But these general statements do not stand up to scrutiny. And crucially the data arise from only a small numbers of cases examined.

In the 15 years covered by the study, only (my italics) 70 cases happened, ie 4.6 a year. Most were HCV+ve, 3 HIV+ve and one HBV+ve; the latter should really be excluded from the list of high risk infection in practice, since all NHS exposure-prone staff have to be successfully vaccinated against HBV and thus this poses no risk. I would contrast this very low cadaver infection rate with what happens at St Thomas' Hospital mortuary, where we see 1-2 HCV and/or HIV+ve cases a week, and have had no problems in performing them safely. Joint compilation of protocols for all eventualities by both anatomical pathology technologists (APTs) and pathologists ensures smooth, safe and efficient practice. I would suggest that the reported reluctance to perform such autopsies depends on unfamiliarity.

With the implementation of universal precautions for all autopsies, the true risks of high risk autopsy practice is minimal. When such infections are common, they do not disrupt the work flows, since all bodies are essentially treated the same, and they do not engender more expense. Let us not forget that in the recent times of good safe working practice, the likelihood of acquiring such an infection at work is vastly less than the risk to life of travelling to and from the workplace.

What are the consequences of reducing open autopsy practice by such minimal invasive techniques, and what are the opportunity costs? The focus here is on persons suspected of drug abuse with HCV or HIV.

Refinement of causes of death I am pleased that in Table 1, autopsies were indeed done to identify the alcoholic ketoacidosis syndrome and co-morbid infections. These variants of toxic pathology cannot be addressed without tissue samples, and preferably open autopsy examination of the relevant organs. But how much other important and/or interesting pathology might be missed by not doing open examinations? Table 1 lists a good number of lesions that may not be seen with CT: heart valve vegetations, tuberculosis (a public health notifiable infection), asthma, and cirrhosis (a disease of public health concern, and not reliably identified by imaging even in the living).

From my own observations I could add three generic scenarios: * The complicated and often critical contribution of co-morbidities, eg chronic lung and heart disease, to death from drug toxicity; it is much easier to evaluate the concepts of borderline toxicity and drug tolerance in individual cases when the whole pathology is known. * The contribution of sepsis from the IV injection habit pe se, both acutely with septic shock, and chronically through amyloidosis and renal failure. * The importance of considering the timing of a drug-related death, such as with evidence of aspiration pneumonitis, and addressing the questions of distressed relatives at inquest - for which there can be much evidence from the autopsy pathology.

Pathology education (or lack of) This is what disturbs me most about these trends towards imaging-only post -mortem examinations. Where and how are we going to teach the next generations of pathologists in the difficult arts of dissection and, even more importantly, histological examination? The current human tissue regulations already impact badly here, and removing yet more case work (drug-related deaths are indeed interesting and have significant internal pathologies that could become unfamiliar) takes away even more opportunity. Whilst some would argue that much of this type of examination is a waste of time, I hold that it provides practice in technique and interpretation, so that when a truly difficult and nuanced case emerges, it can be addressed with experience and reason.

Research opportunities Coronial autopsies are not intended for research but, basically, to determine whether a cause of death is natural, and an inquest may be dispensed with, or actually or potentially unnatural and so needs more investigations and inquiry. That said, because they provide >95% of adult autopsy work in the UK, they inevitably have a surveillance and potential research role. The epidemiology of diseases, including infections, changes constantly, and the autopsy provides one mode of monitoring and reporting on this.

The prime common example (I omit transmissible spongiform encephalopathies deliberately) is HIV disease. Much of what we know of the clinico-pathological cadence of HIV disease and results of new treatments (beneficial and adverse) comes from autopsy work. And it is published as such, although the commissioning coroners are probably not aware of that.

Coronial autopsies make a significant contribution to our understanding of cardiovascular disease in HIV-infected persons. Those not familiar with HIV may not realise the large clinical research, treatment and pharma interest into whether HIV per se and/or its anti-retroviral therapies do, or do not, activate endothelial cells and so augment arteriosclerosis, affecting the heart and brain in particular. HIV-infected suspected drug abusers thus contain within themselves at least two interesting pathological aspects (what is HIV doing and what are the drugs doing to that person), where a full autopsy can provide unique and cumulative evidence, to the ultimate benefit of public health.

Requirement for minimal invasive post-mortem examinations As Fryer et al state, the minimal invasive system requires two robust processes in place. First rapid toxicology, and they indicate one week as satisfactory. I would argue that this is not fast enough, since bodies do decompose even whilst refrigerated and important histopathological information is lost. More practically, in London, none of the laboratories offering services performs even that fast. That should be remediable, if the paymasters (the coroners) exercised their power to force the laboratories to turn over tests within, say, 3 working days.

Secondly, available imaging, particularly CT scanning. This is in practice impossible without proper funding; I pass over the availability of interested pathologists. At present, such non-forensic imaging is funded from now rather old government grants, or from individual initiatives such as jewish or moslem communities, or even interested radiologists with some surplus monies in their educational funds. But these are not appropriate for a nation-wide roll-out of cadaveric imaging - whether as replacement or (I would argue) as adjunct to open autopsy. These post-mortem examinations are done at the behest of coroners, and unfortunately they are not centrally but locally funded, with all that implies for variation in service provision.

So is there a future plan? The NHS has recently issued a large post- consultation document on cadaveric imaging {ref}, written by Prof Guy Rutty in Leicester and colleagues, with input from many other relevant specialities. I do encourage all autopsy-active pathologists (and coroners) to read it.

It provides the first realistic estimates of the actual costs of autopsies, with or without imaging costs, which alone make enlightening and disturbing reading for those involved in the economics of autopsy practice. But its main plank is the plan for future mortuary provision in England. Essentially, it is proposed that all mortuaries have attached dedicated CT scanners; that there need be only 30 such facilities in England (only 3 in London, the rest outside). And that all bodies are scanned prior to autopsy. The optimal funding for such an integrated pathology-radiology service is central government, not local source.

Conclusion There is much controversial material in this NHS document to discuss, but I certainly endorse the significant reduction of active mortuaries, with provision of imaging facilities on-site, and the resulting concentration of expertise in such places. As well as cases I still perform myself, I review many autopsies done by others and am frequently disturbed by their suboptimal or frankly dreadful quality. It is inevitable that experience, insight and - crucially - constant audit by, and consultation with, peers does sharpen and maintain practice standards. And so with autopsies: we should be doing them as a speciality interest practice, with similarly interested colleagues, in centres that do a lot of them very well. Which brings me back to the start of this Comment: in such facilities, 'high risk infections' will pose no problems for practitioners, who will be very familiar with them and their wrinkles. So isolating that category of cases for a qualitatively different approach to post-mortem examination from all the other cases will not be necessary.

Whether the NHS plan is rolled out as proposed, or through other exigencies the number of active mortuaries declines, the end result of fewer, but properly specialised facilities is appropriate. Pre-examination imaging will find it right place and 'high risk' cases will be optimally prosected.

Prof Sebastian Lucas Dept of Histopathology St Thomas' Hospital London SE1, UK Sebastian.lucas@kcl.ac.uk 28th Jan 2013

Reference "Can Cross-Sectional Imaging as an Adjunct and/or Alternative to the Invasive Autopsy be Implemented within the NHS?" Report from the NHS Implementation Sub-Group of the Department of Health Post Mortem, Forensic and Disaster Imaging Group (PMFDI). October 2012. The document can be downloaded from the East Midlands Forensic Pathology Unit. The full website address is: http://www2.le.ac.uk/departments/emfpu/Can%20Cross- Sectional%20Imaging%20as%20an%20Adjunct%20and- or%20Alternative%20to%20the%20Invasive%20Autopsy%20be%20Implemented%20within%20the%20NHS%20 -%20FINAL.pdf

Conflict of Interest:

None declared

Dear Editor,

I read with interest the recent original article entitled 'Mucosal large cell neuroendocrine carcinoma of the head and neck regions in Japanese patients: a distinct clinicopathological entity' by Kusafuka et al.[1] The patient (Case 2) in this article had been previously reported.[2] I note some discrepancies between these two papers. In this original article, the patient (Case 2) was a 65-year-old man. The mitotic rate of this tumor was not specified, but >15 per 10 high power fields. Immunohistochemically, the tumor cells were focally positive for chromogranin-A, but negative for synaptophysin. The Ki-67 labeling index was 90.8%[1]. In the case report[2], this patient was a 79-year-old man. The mitotic rate was 13 per 10 high-power fields. The tumor cells were focally positive for synaptophysin (shown in Figure 3c), but negative for chromogranin-A. The Ki-67 labeling index was 82%.

There are also some discrepancies between this original article and their recent review article[3]. In the review[3], they mentioned that four of their eight mucosal large cell neuroendocrine carcinomas arose in the larynx (supraglottis). The follow up periods were 12-96 months, and only one patient died of disease. The authors concluded that the Japanese cases of mucosal large cell neuroendocrine carcinoma had a better prognosis than that reported in the literature. Three cases were immunopositive for thyroid transcription factor-1. But in this original article[1], four tumors occurred at the larynx, with three at the supraglottis and one at the infraglottis. Three patients died of disease, and one patient died of another disease (shown in Table 2), which led to the conclusion that their Japanese series of large cell neuroendocrine carcinoma also indicated a relatively poor prognosis as that reported in the literature. The follow up periods were 15-90 months (shown in Table 2). Only two cases were immunopositive for thyroid transcription factor-1.

Discrepancies were also present in this original article[1]. In the section of Results, the age of the eight patients ranged from 52 to 75 years (mean 64.6 years). In the section of Discussion, they stated all the cases in the present series were aged >65 years and the follow up periods of the patients that have been disease-free were 24-108 months. But in Table 2, the age ranged from 52 to 74 years and four patients were alive without disease (31 months, 18 months, 24 months and 90 months after surgery, respectively).

In conclusion, there are many inconsistent data in these articles concerning mucosal large cell neuroendocrine carcinoma by Kusafuka et al.[1-3] They should verify the original data and make a correction.

Sincerely, Shaodong Yang. Department of Oral Histopathology, Hainan Medical College, 3 Xueyuan Road, Longhua District, Haikou, China.

REFERENCES 1. Kusafuka K, Abe M, Iida Y, et al. Mucosal large cell neuroendocrine carcinoma of the head and neck regions in Japanese patients: a distinct clinicopathological entity. J Clin Pathol 2012;65:704-9. 2. Kusafuka K, Asano R, Kamijo T, et al. Large cell neuroendocrine carcinoma of the tongue base: case report of an unusual location with immunohistochemical analysis. Int J Oral Maxillofac Surg 2009;38:296-9. 3. Kusafuka K, Ferlito A, Lewis JS Jr, et al. Large cell neuroendocrine carcinoma of the head and neck. Oral Oncol 2012;48:211-5.

Conflict of Interest:

None declared

Dear Editor,

We read this article with great interest and would like to share our own similar experiences in support of this growing evidence base. Our department has the added complexity of being one of the UK ST1 training schools, with between ten and fifteen ST1 - ST5 trainees per year. We have trained and developed a senior Biomedical Scientist (BMS) in all specimen dissections who has gained the RCPath Diploma of Expert Practice in Histological Dissection and been appointed as an Advanced Practitioner (AP). This post covers all surgical specialties, and the AP currently dissects the majority of specimens, both simple and complex, as and when appropriate knowledge and experience has been gained. In terms of colorectal specimens, all types are dissected by the AP, including Extralevator Abdomino-Perineal Excision (ELAPE) specimens where the levator ani and/or coccyx are also resected.

Macroscopy

In terms of macroscopic assessment we follow a similar defined protocol to that of Sanders et al, with triage of each colorectal cancer specimen before dissection. Included in this discussion is most appropriate block selection. This process occurs for both the AP and the trainees, with the AP following the same pathway as that of the trainees. The AP and trainee take similar numbers of blocks, as appropriate to the case, in keeping with the departmental protocol (at least four blocks required). A macroscopic proforma is used to record data items during dissection. This is used by all staff, regardless of grade, and is seen as a method of ensuring that the minimum data items are recorded. Evidence has shown there to be more chance of identifying extramural vascular invasion (EVI) where additional tumour blocks are sampled [1]. In the study by Sanders et al it would be interesting to know how often EVI was identified where additional tumour blocks were taken by the BMS, as this may not be an appropriate criticism of practice.

Audit

We recently performed an audit to assess lymph node harvesting. This showed the average lymph node harvest of the AP was twenty-five percent higher that that of her histopathologist colleagues (consultant and trainee) (20 vs 15). All staff groups achieve the RCPath requirement of a harvest of twelve lymph nodes per case [2]. These findings were statistically significant (p=<0.001) and were presented at a national meeting [3]. The AP was less likely to perform resampling for additional lymph nodes (8.1% vs 16.6%). This was attributed to the level of training and experience of the AP in comparison to some trainee pathologists. In Sanders et al study the mean numbers of lymph nodes harvested by medical staff are similar to our own, but those harvested by the BMS are lower. This may be related to the experience of the BMS, as a similar decrease is seen with the least experienced trainees in our laboratory. Time spent on dissection may be a factor as our AP takes on average forty-five to sixty minutes per case.

Microscopy

As part of the Histopathology team our AP routinely reviews the slides generated from her colorectal cancer dissections. Reports are generated using a microscopic minimum dataset proforma and a paper copy of this is initially completed by the AP for joint review with the consultant before generation of the final report and authorisation by the consultant. This pathway is identical to that of the trainees within our department.

Development of scientific staff

Our AP has a specific interest in gastrointestinal pathology and is currently leading a research study around colorectal cancer diagnostics as part of a Professional Doctorate in Biomedical Science. This follows the Modernising Scientific Careers (MSC) programme which aims to provide the most appropriate developmental opportunities to high achieving scientists in their field [4]. Achieving the RCPath Specialist Diploma in lower gastrointestinal histopathology is also a current aim for our AP. To improve the knowledge base surrounding colorectal cancer management, she also regularly attends the local colorectal cancer multidisciplinary team meeting with one of the specialist consultants. This has the additional benefits in that it publicises the role of the AP within the team, and also provides an essential feedback mechanism on her performance. We feel that the performance of the AP is similar to that of an experienced registrar, and that she should be treated in the same way.

Benefits

The benefits to us are multiple. Consultant histopathologists benefit from more time to perform other important duties and the ability to delegate dissection training where appropriate. We would argue that in many cases reporting is quicker with the AP than with a trainee, as specimens are less likely to require resampling. Trainee histopathologists benefit from the additional dissection training, supervision and an additional source of advice. Due to the protocol driven nature of the AP's practice, she is able to provide the trainees with examples of agreed best practice within the department. APs benefit from career development which was previously unavailable, with the potential of eventual consultant healthcare scientist status when at doctoral level and with appropriate experience [4]. Patients benefit from the knowledge that dissection practice is standardised and that they are receiving the correct treatment based on appropriate staging.

In conclusion, we feel that there is an important clinical role for scientists within the histopathology team. Rather than simply working on the laboratory based aspects of the service, high achievers within this staff group should be utilised more appropriately. This may be in specimen dissection and microscopy, but could equally be within immunohistochemistry or molecular pathology. We agree that the evidence base needs to be improved in order to support further development within this area.

Conflict of Interest:

None declared

Dear Editor,

A very important issue was assessed by Yang et al in their outstanding study recently published by this journal1. As pointed out by the authors, distant metastasis is the most preoccupant complication of differentiated thyroid carcinoma (DTC) and a very anguishing therapeutic challenge for the attending physician. A series of studies have been trying to establish a molecular pattern able to predict more aggressive follicular cell behavior. One of the most promissory markers integrating this molecular pattern is the expression of neural cell adhesion molecules (NCAM).

Yang et al studied a series of 365 surgical cases of thyroid disease - 214 DTC and 151 benign lesions. Immunohistochemistry showed that most benign lesions presented NCAM expression, whereas a significant proportion of DTC lost completely or showed a reduced NCAM expression, which confirms previous results suggesting that NCAM could be a diagnostic marker of DTC 2. We also studied NCAM expression in a series of 527 surgical cases of thyroid tissues - 395 DTC (343 papillary thyroid carcinomas and 52 follicular carcinomas) and 132 nonmalignant thyroid tissues (18 normal thyroids, 58 goiters and 56 adenomas). One hundred fifty-three of our patients presented metastasis at diagnostic and 58 developed distant metastasis during a follow-up of 12-298 months (43.50?33.29 months), Mo=21 months. NCAM expression was evaluated by immunohistochemistry and the same technique used by Yang et al, but with anti-NCAM monoclonal 123C3 clone antibody (DAKO- Carpenteria, CA, USA). We also considered NCAM positive those cases with NCAM expression in more than 30% of tumor cells. Fisher's exact test showed total loss or reduction of NCAM expression in 74.65% of DTC cases, while a significant portion (52.73%) of benign lesions were positive for NCAM (p< 0.0001). However, NCAM expression was not able to predict malignancy due to low sensibility (25.35%) and low specificity (47.27%), suggesting that NCAM alone is not a useful diagnostic marker.

In addition, Yang et al found that persistent NCAM expression in DTC is associated with a higher rate of metastasis. In our cohort, NCAM expression was not correlated with the presence of metastasis at diagnosis (p=0.4506), neither to tumor size (p=0.3814) nor to extrathyroid invasion (p=0.9855), multifocality (p=0.2747) or pTNM stage (p=0.6928). A log-rank test failed to show NCAM expression as a prognostic marker of relapse-free survival (p=0.8846). Nevertheless, NCAM positivity was more frequent in encapsulated tumors (37.78%) than in nonencapsulated tumors (20.62%; p=0.0399), suggesting that the peritumoral fibrotic reaction is associated with NCAM expression. In fact, 51.11% of our NCAM positive cases presented concurrent chronic lymphocytic thyroiditis (CLT), while only 25.89% of NCAM negative cases presented concurrent CLT (p=0.0045). We also evaluated the presence of tumor infiltrating lymphocytes (TIL) in DTC specimens by a routine HE staining. We found that NCAM expression was associated with the presence of TIL (p=0.0427). In order to distinguish TIL subsets, we performed immunohistochemical analysis using classical immune cell markers. We observed that NCAM expression was associated with the presence of CD4+ lymphocytes (p=0.0477), CD8+ lymphocytes (p=0.0015), CD20+ lymphocytes (p=0.0284) and FoxP3+ lymphocytes (p=0.0024). Interestingly, most NCAM negative cases (80.10%) were also negative for sodium/iodine symporter (NIS) protein immunohistochemical expression, whereas 71.43% of NCAM positive cases were positive for NIS as well (p< 0.0001), suggesting that NCAM could boost immunogenicity in DTC. These results suggest that NCAM expression is engaged in the antitumor immune response. However, the outcome of patients is not modified by NCAM expression, perhaps because an appropriate management of DTC patient is the most important and modifiable prognostic factor, impeding the natural course of malignancy.

The differences between Yang results and our data could be related to different population backgrounds, which are thought to affect antitumor immunity in DTC 3. Since the tumorigenic process is a complex biological system in which multiple molecular interactions may occur, minimum genetic differences in populations might affect dramatically the obtained results. An antitumor effect of NCAM may be expected in cases presenting genetic background that facilitates antitumor immune defense 3-4. We also cannot exclude that the different antibody used may lead to different results. In addition, it is worthy noting that NCAM may be engaged in pleiotropic functions in tumor progression, making the interpretation of NCAM expression a difficult task. More studies are warranted to understand the functional biologic role of NCAM expression in DTC tumors. Unfortunately, our data do not support the conclusion of Yang et al that NCAM expression in well differentiated thyroid carcinoma is an indicator for a higher risk of distant metastasis.

Sincerely,

Lucas Leite Cunha1, Elaine Cristina Morari2, Suely Nonogaki3, Fernando Augusto Soares4, Jose Vassallo5, and Laura Sterian Ward1.

1Laboratory of Cancer Molecular Genetics, Faculty of Medical Sciences - University of Campinas (Unicamp). 126 Tessalia Vieira de Camargo Street, Campinas, SP, Brazil.

2Department of biological sciences and health- State University of Roraima. 231, Sete de Setembro Street, Boa Vista, Roraima, Brazil.

3. Adolfo Lutz Institute. 355, Doutor Arnaldo Avenue, S?o Paulo, Brazil.

4Department of Pathology, A. C. Camargo Cancer Hospital. 211 Antonio Prudente Street, S?o Paulo, SP, Brazil.

5Laboratory of Investigative and Molecular Pathology (Ciped), Faculty of Medical Sciences - University of Campinas (Unicamp). 126, Tessalia Vieira de Camargo Street, Campinas, SP, Brazil.

REFERENCES

1. Yang AH, Chen JY, Lee CH. Expression of NCAM and OCIAD1 in well- differentiated thyroid carcinoma: correlation with the risk of distant metastasis. J Clin Pathol 2011;

2. El Demellawy D, Nasr AL, Babay S, Alowami S. Diagnostic utility of CD56 immunohistochemistry in papillary carcinoma of the thyroid. Pathol Res Pract 2009;205:303-9.

3. Cunha LL, Tincani AJ, Assumpcao LV, Soares FA, Vassallo J, Ward LS. Interleukin-10 but not interleukin-18 may be associated with the immune response against well-differentiated thyroid cancer. Clinics (Sao Paulo) 2011;66:1203-8.

4. Scarpino S, Di Napoli A, Melotti F, Talerico C, Cancrini A, Ruco L. Papillary carcinoma of the thyroid: low expression of NCAM (CD56) is associated with downregulation of VEGF-D production by tumour cells. J Pathol 2007;212:411-9.

Conflict of Interest:

None declared