An article in this issue of Medical Teacher reports the demise of the ‘Breakfast Club’, a voluntary, student-centred, casebased, self-directed learning system, based on attendance at autopsies, and involving learning about basic science, deductive reasoning, clinical correlation, peer evaluation, medical history and respect for the dead (O’Grady, 2004). The reason for the cessation of this educational activity was the legally binding decision to prohibit student access to coronial autopsies in Auckland, New Zealand. Why was student access prohibited? It comes down to the absence of consent for the use of the body for educational purposes (O’Grady, 2003). Similar concerns about the lack of specific consent for teaching have stopped regular ‘live’ student autopsy teaching in many UK medical schools including mine, as the number of autopsies being requested with specific permission for use in teaching is too small to permit regular timetabling. When one adds to this the overall reduction in the autopsy rate in teaching hospitals in the United Kingdom over the past few years (Wright, 2002), due in part to issues over consent and also due to recent highly publicized concerns over organ retention, we are in danger of losing both the ultimate audit of clinical diagnosis and treatment, and the means of adequately training future generations of doctors. Most sensible individuals agree that the process of obtaining informed consent or authorization for the use of human tissues for education or research needs to be improved, clarified and properly implemented. However, if it was appreciated that requiring formal additional consent for teaching, on an autopsy that was being performed anyway, would result in significantly less autopsy teaching (Wright, 2002; O’Grady, 2003, 2004) and consequently less welleducated doctors (evidence presented below), I wonder how many people would continue to insist on this? There is much that is good in the new medical curricula being taught in UK medical schools. The integration of teaching of the major disciplines in a systematic way, the development of clinical skills centres, the imaginative use of simulators and simulated patients, and the emphasis on selfdirected learning, on communication skills and on becoming a lifelong learner are all steps towards producing better doctors. But practising good medicine does depend fundamentally on a knowledge and understanding of how the body works or malfunctions. I would contend that one of the best ways of learning this is through autopsy-centred discussion. I base this view on a number of pieces of evidence and experience. In the article in this issue we are informed that the members of the Breakfast Club placed consistently high in their classes. In one of the most enjoyable and productive phases of my career, I was Senior Lecturer in Pathology in St Bartholomew’s Hospital in London. Here the tradition was to present the morning’s autopsy findings to a gallery of medical students and doctors at 1.30 p.m. daily. During these demonstrations I was regularly challenged by physicians and students to explain findings or lack of findings in the autopsy. I took equal satisfaction from confirming or refuting clinical diagnoses, and sometimes I could not give a definite answer. Many medical students from that period, some of whom are now distinguished professors, comment to me on how valuable and educational they found these demonstrations. One former student told me recently that he remembers a particularly spirited exchange between me and one of the most driven senior physicians as ‘one of the highlights of his time as a medical student’. The physician had given the clinical history and concluded by saying something like ‘‘the patient then seemed to lose the will to live and turned his face to the wall and died’’. As I looked at the massive pulmonary embolism that occluded both pulmonary arteries in the lungs on the demonstration bench in front of me, I simply said that I thought we could be slightly more precise than that about the cause of death. The students obviously appreciated this perspective and probably all remember and understand the importance of pulmonary embolism and its subtleties. Another experience from that period involved the same irascible senior physician who was not very happy when, at the teaching autopsy demonstration, I could not pinpoint the cause of the deceased’s illness or death (the deceased had been a complete clinical mystery in life)—I could only note at the demonstration that most organs appeared swollen and congested, but I did say I would sample tissues widely for histology and store some fresh tissue in the freezer (not legal now without specific consent). I still remember the first section that I looked at under the microscope two days later. It was brain showing virtually every single capillary packed with malignant cells, but no tumour mass. Sections of every organ showed the same picture, and with immunohistochemistry on the paraffin sections and gene rearrangement studies on the fresh frozen tissue, we were able to diagnose this unequivocally as an angiotropic large B-cell lymphoma, thus

We previously observed, in decalcificated bone specimens, intraosseous crystal deposits with morphological and analytical similarity to calcium pyrophosphate dihydrate. We have now been able, by a combination of more detailed morphological studies of these and similar cases, and by infrared spectroscopy in three cases, to show that this is, in fact deposition of the secondary calcium salts brushite and monetite, occurring as an artefact during formic acid decalcification. Our earlier postulate of bone as an additional primary crystallization site for calcium pyrophosphate dihydrate is effectively disproved. This artefact deserves wider recognition.

p53 gene mutation appears to play an important role in the development of systemic lymphoma, and may be associated with tumour progression. Its role in cutaneous lymphoma is currently unknown. We examined p53 expression in 55 biopsies of cutaneous lymphoma, including patch-, plaque- and tumour-stage mycosis fungoides (MF), T- and B-cell lymphoma and lymphomatoid papulosis. Strong, homogeneous p53 expression, thought to correlate most closely with p53 gene mutation, was seen in only three cases; in a plaque and tumour from a patient with tumour-stage MF, in plaque-stage MF in a patient without tumours, and in one case of CD30+ large-cell anaplastic lymphoma. These data suggest that p53 gene mutation is not a critical step in the development of the majority of primary cutaneous lymphomas.

The second British Stomach Cancer Group trial was a prospective randomised controlled trial of adjuvant radiotherapy or cytotoxic chemotherapy after gastrectomy for adenocarcinoma. It recruited between 1981 and 1986. No survival advantage has been demonstrated for the patients receiving either type of adjuvant therapy compared with those undergoing surgery alone. We report on 436 patients randomised into the trial together with 203 patients, who did not fulfil the trial criteria, referred to the trial. A univariate (log-rank) analysis of pathological factors obtained from the local referring centres showed that tumour size, macroscopic type, number os sites involved, depth of invasion, involvement of resection lines and lymph nodes and histological grade were significant determinants of survival. Histological review by two experienced histopathologists found that the Lauren classification and histological grade, but not the Ming classification, were significant prognostic factors. The degree of lymphocytic and eosinophilic infiltration and presence of dysplasia assessed by one of the pathologists showed a significant correlation with survival. However, inter-observer correlation for these histological parameters and grade was poor. Multivariate analysis identified only depth of invasion, resection line and nodal involvement as significant independent pathological variables influencing survival. This study confirms the need for expert preparation of the resected specimen to obtain the important information on depth of invasion and nodal status and also reveals some variation in histological assessment, particularly grading, in gastric carcinoma.

Antibodies to proliferating cell nuclear antigen (PCNA) and the MIB-1 antibody to the Ki-67 antigen were titrated to optimize identification of proliferating cells in formalin-fixed paraffin-embedded tissue from a series of 40 human breast carcinomas. Cell culture studies have previously demonstrated that immunostaining for both PCNA and the Ki-67 antigen produces strong granular nuclear staining during S phase. PC10, other anti-PCNA antibodies (PC2, PC5, PC8 and 19F4) and MIB-1 were used at the minimum dilution which allowed a clear distinction between cells with strong and weak staining. With the anti-PCNA antibodies, nickel-cobalt enhancement of the reaction product was found to augment the granular nature of nuclear staining, corresponding more closely to patterns observed in cell culture studies. No enhancement was found to be necessary for MIB-1. The labelling indices of all these antibodies were compared with S phase fraction (SPF) obtained by DNA flow cytometry in the same cases. The PC10 labelling indices which included only strongly stained cells correlated well with SPF, but counting all strongly and weakly stained cells showed a poor correlation. With MIB-1, counting strongly stained as well as all stained cells produced labelling indices which correlated well with SPF, the former tending to be lower and the latter higher. None of the other anti-PCNA antibodies showed any advantage in application over PC10. Thus, PC10 and MIB-1, applied with care, can be correlated with S phase fraction in paraffin processed tissue sections of breast carcinomas.

Aims-To describe the birefringent saponified fatty acid crystalloids seen in pancreatic fat necrosis.Methods-A histological review, including polarising microscopy, of three cases of subacute or subclinical acute pancreatitis was performed. Histochemical analysis using Nile blue sulphate for lipid, Holczinger's copper rubeanate for fatty acids, and Alizarin Red S for calcium was performed in one case. Scanning electron microscopy and x-ray energy dispersive spectroscopic microanalysis were performed in two cases. Necropsy pancreatic tissue, surgical archival tissue from cases of pancreatitis, and pancreatic and adipose tissue permitted to autolyse together in the laboratory, were also examined. The autolysed tissue was also examined histochemically. Stained and unstained sections were mounted in DPX and Canada balsam. Surgical material showing traumatic fat necrosis was reviewed.Results-In each of the three cases there were subtle clues to subclinical pancreatitis. In neither surgical case was the true nature of the mass apparent to the operator. Histological analysis in all cases showed ghost adipocytes containing numerous polarising crystalloids, as well as some basophilic debris. Microanalysis showed calcium but no other substantial heavy element signals. Histochemical analysis showed a labile, polar, acidic lipid and the crystalloids behaved as calcium salts of free fatty acid, The crystalloids were not seen in archival material mounted in Canada balsam. No crystalloids were seen in traumatic fat necrosis. Conclusions-Little recognised, strongly birefringent, saponified free fatty acid crystalloids occurring in pancreatic fat necrosis may survive routine processing, and can point to the origin of obscure mesenteric masses related to subclinical pancreatitis.

If skilled histopathologists disagree over the same biopsy specimen, at least one must have an incorrect interpretation. Thus, disagreement is associated with, although not the cause of, diagnostic error. The present study aimed to determine the magnitude of variation among 10 observers with a special interest in gastrointestinal histopathology. They independently interpreted the same biopsy specimens for morphological features which may discriminate between patients with Crohn's disease and ulcerative colitis and normal subjects. Thirty of 41 features had agreement measures significantly better than expected by chance (p < 0.05). The range of agreement in the 45 observer pairs over the final diagnosis was 65-76%. There was good agreement in discriminating between normal slides and those showing confirmed inflammatory bowel disease. For normal slides, however, the term nonspecific inflammation was often applied and without any consistency. In addition, true Crohn's disease slides were often and consistently thought to be ulcerative colitis. Having identified 11 important discriminatory morphological features, two multiple regression analyses were then carried out to produce a scoring system for inflammatory bowel disease. These results suggest there is considerable room for improvement in the reliability of colonic biopsy specimen interpretation and that this could probably be achieved using more exact definitions of morphological features and diseases.

BACKGROUND:The histologic grading of the deep invasive margin of oral squamous cell carcinoma recently has been shown to have prognostic value, but previous series have not been homogeneous enough to allow grading parameters to be assessed individually.METHODS:Forty-seven small lingual carcinomas limited to the lateral border of the tongue and treated by radiotherapy were graded histologically at their deep invasive front. Clinical and grading parameters were correlated by statistical tests performed by permutational techniques.RESULTS:Carcinoma recurred locally in 6 patients, and metastases developed in 19. Local recurrence correlated with Broders' grade (P = 0.0143), keratinization (P = 0.017) and pattern of invasion (P = 0.0195). Metastasis had a highly significant correlation with Broders' grade (P < 0.001), pattern of invasion (P < 0.001), and invasive front grading total score (P < 0.001). Seven of 8 carcinomas with diffuse infiltrating patterns metastasised, whereas only 4 of 25 with large islands or a broad infiltrating pattern metastasized.CONCLUSIONS:The usefulness of the deep invasive front grading system for small lingual carcinoma was demonstrated. The pattern of invasion was the component of the grading system that had the closest correlation with metastasis and recurrence in this type of carcinoma.

A 29-year-old man presented with nodular skin lesions localized to areas of navy-blue pigmentation within a tattoo. Light microscopy demonstrated well-defined epithelioid granulomata in close relation to blue and black pigment. Although the patient was asymptomatic, a chest X-ray showed bilateral pulmonary shadowing, and histology of a transbronchial biopsy specimen showed features compatible with a diagnosis of sarcoidosis. X-ray energy dispersive spectroscopy identified copper and titanium in the tattoo pigment. These elements were not found in tissue from the lung biopsy. The cutaneous eruption resolved with oral steroid therapy. Our observations suggest that the granulomatous reaction in the tattoo was a manifestation of sarcoidosis, rather than a specific reaction to pigment.

We describe a case of primary T cell lymphoma of the liver developing in a patient with Felty's syndrome (FS). We discuss the possible relationship of the two conditions with particular reference to liver disease in FS, and the role of the T cell in RA.

The pattern of c‐erbB‐3 gene product was studied in 91 advanced gastric carcinomas, adjacent hyperplastic mucosa, intestinal metaplasia and dysplasia and in normal controls, using immunohistochemistry in archival material. All tumours showed positive c‐erbB‐3 staining in both cytoplasm and membrane. No significant differences of expression were observed between intestinal and diffuse‐type carcinomas or any other clinical parameters. Of interest is the expression in the adjacent mucosa, which is extensive, cytoplasmic, and of lower intensity than in the tumours. Further studies are currently being carried out to clarify the role of this protein in tumour behaviour and gastric carcinogenesis.

Immunohistochemical staining was carried out on a spectrum of normal, hyperplastic and malignant endometrial curettings, for proliferating cell nuclear antigen--PCNA (using the monoclonal antibody PC10) and for abnormally stabilized p53 (using the polyclonal antibody CM-1). The mean proportion of glandular epithelial cells showing PCNA immunoreactivity was significantly lower in atypical hyperplasia/intra-endometrial adenocarcinoma than in invasive adenocarcinoma, but the degree of overlap between the cases was such that this was not considered to be of diagnostic value. p53 immunoreactivity was detected in 47% of invasive adenocarcinomas and in a much smaller proportion of endometria showing simple hyperplasia and atypical hyperplasia, but staining was only focal in the last two conditions. The majority of p53-positive invasive adenocarcinomas had a large proportion of glandular epithelial cells expressing PCNA, but a significant number of p53-negative cases also had a high PC10 index. This suggests that, in endometrial neoplasia, there is not a simple relationship between abnormally stabilized p53 and PCNA expression.