MRCGP pass rate by medical school and region of postgraduate training. Royal College of General Practitioners.

Abstract

Department of Oncology,\rQueen's University of\rBelfast, Belfast BT9 7BL\rW S Lowry, professor\rRJ Atkinson, senior lecturer\rCorrespondence to:\rProfessor Lowry.\rBMJ 1993;307:542\rCorrespondence to:\rMr Wakeford.\rBMJ 1993;307:542-3\rTumour suppressor genes and\rrisk ofmetastasis in ovarian\rcancer\rWS Lowry, RJ Atkinson\rThe silent onset of ovarian cancer often leads to\rmetastatic spread before diagnosis can be made. The\rcause remains unknown, but molecular studies are\rbeginning to reveal some mechanisms. Inactivation of\rtumour suppressor genes is thought to be responsible\rfor initiating many forms of cancer. Studies of allele\rloss indicate areas of genome where inactivation may\roccur. We recently examined the relation between the\rclinical stage of epithelial ovarian cancer and allele loss\rat three loci on chromosome 17: progressive loss was\rdetected with advancing stages of disease.'\rThe clinical stage of a tumour defines the extent of\rthe disease and is the best determinant of prognosis for\rmost forms of cancer. It depends on the time elapsed\rbefore diagnosis, hence the value placed on early\rdetection. The histopathological grade, however,\rbetter describes the intrinsic nature of the tumour: well\rdifferentiated tumours are slow growing whereas undifferentiated,\ranaplastic tumours metastasise rapidly.\rThus poorly differentiated tumours tend to be\rdiagnosed at a more advanced stage. At the molecular\rlevel one might anticipate less damage to genetic\rmaterial with well differentiated tumours.\rMethods and results\rWe collected 52 fresh specimens of ovarian tumours,\rofwhich 38 satisfied the pathological criteria of invasive\rmalignant disease. Complete clinical and laboratory\rdata were available for 33 of the cases, but three were\rhomozygous and thus uninformative. The remaining\r30 cases were informative at one or more of the three\rloci examined on chromosome 17 (two polymorphic\rloci (YNZ22.2 and BHP53) on l7pl3 and one at the\rTHH59 locus (17q23-qer)). DNA was extracted from\rfresh blood samples and was analysed by Southem blot\rhybridisation. The tumours were classified histologically\rinto well, moderately, and poorly differentiated\rlesions according to grades I, II, and III respectively.\rThe greatest allele loss was seen in grade III tumours\rat all three loci, with the highest loss of over 90% at\rTHH59) (table). Allele loss was least in grade I\rtumours, with no loss at 17p. Results for grade III\rtumours were compared with combined results for\rgrade I and II lesions at each locus with Fisher's test of\rexact probability (two tailed): significantly greater\rallele loss was associated with grade III lesions at\rTHH59 (11/12 v 2/10, p=0 002) and YNZ22 (10/12 v\rAllele loss at three loci on chromosome 17 by histopathological grade of\rtumour among 30 cases of ovarian cancer. Values are proportions of\rthose cases that were informative at each ofthe loci\rGrade oftumour\rLoci on\rchromosome 17 III II I\rTHH59 11/12 1/3 1/7\rYNZ22 10/12 1/6 0/4\rp53 3/5 1/2 0/2\rTotal 24/29 3/11 1/13\r1/10, p=0002), but the difference between the small\rnumber of observations for p53 was not significant\r(3/5 v 1/4, p= 0 7 14). When the results for the three loci\rwere combined the difference between grade III lesions\rand grade I and II lesions was highly significant\r(24/29 v 4/24; X2=2044, df=l; p<0 001) (95%\rconfidence interval for the difference 45% to 87%).\rComment\rThe histological grade of a tumour is difficult to\rdefine and measure accurately, and it may deteriorate\rwith advancing malignancy as cells undergo multiple\rabnormal divisions. Despite such criticisms, however,\rmost clinical and pathological studies agree that grading\ris related to survival.2 In our previous analysis of\rtumour stages, allele loss on 17p was seen as a late step\rin the progression of ovarian malignancy. This seems\rto be ubiquitous in many tumours and is probably\rrelated to inactivation of the p53 tumour suppressor\rgene. Allele loss at THH59 is an earlier step and\rappears to be specific to ovarian cancer.\rPrognostic factors can be analysed more precisely\rwhen staging and grading are considered together.\rEvidence suggests that loss of a putative tumour\rsuppressor gene at THH59 is significantly higher in\rundifferentiated cancers. These rapidly growing\rtumours have a greater potential for metastatic spread\reven in early stage disease. The highly significant\rdifference in the cumulative observations strongly\rsuggest that anaplastic tumours have lost the entire\rchromosome. The degree of malignancy of the tumour\rthus depends on the extent of underlying molecular\rdamage.\rWe thank our colleagues in gynaecology and pathology; Dr\rHilary Russell, who initiated the earlier laboratory studies;\rand Mrs Patricia White for technical assistance.\r1 Eccles DM, Russell SEH, Haites NE, Atkinson R, Bell DW, Gruber L, et al.\rEarly loss of heterozygosity on 17q in ovarian cancer. Oncogene 1992;7:\r2069-72.\r2 Anderson MC. In: Sharp F, Soutter WP, eds. Ovarian cancers-the way ahead\rSection 2. Chichester: Wiley, 1987:63-4.\r(Accepted 0May 1993)\rMRCGP pass rate by medical\rschool and region ofpostgraduate\rtraining\rRichard Wakeford, John Foulkes,\rChris McManus, Lesley Southgate\rInitiatives to measure and enhance the quality of\rhigher education for both undergraduates and postgraduates\r(academic audit, teaching quality assessment)\rwill shortly start to impact on medical schools in the\rUnited Kingdom.\rWe believe that it would be helpful to consumers of\rmedical education (prospective students), to its providers\r(medical schools and their curriculum designers,\rpostgraduate training schemes), and to its customers\r(medical care providers) if comparative output data\ron medical schools and postgraduate training were\ravailable. Without a national medical qualifying\rexamination this is probably possible only by examining\rthe performance of candidates for the major postgraduate\rexaminations of the royal colleges.\rMethods and results\rTo start such a process we analysed the pass-fail\rresult of seven recent diets (December 1988-December\r1991) of the membership examination of the Royal\rCollege of General Practitioners (MRCGP) by clinical\r542 BMJ voLumE 307 28AUGUST 1993\rMRCGPpass rates by United Kingdom medical school of training and region ofpostgraduate training\rNo taking No taking\rMedical School exam % Pass Region exam % Pass\rOxford 70 94-3 Devon and Cornwall 109 94-5\rBristol 125 92-8 Overseas 18 94-4\rNottingham 124 91.1 NorthemrIreland 141 92-2\rSouthampton 98 90-8 East Anglia 108 90 7\rCambridge 74 90-5 North West Thames 118 89-8\rRoyal Free 63 90 5 Trent 206 88-8\rNewcastle 138 89-9 Oxford 152 88-8\rKing's 48 89-6 Northem 225 88-4\rSt Mary's 77 89-6 Avon and Somerset 138 88-4\rBirmingham 120 89-2 North East Scotland 51 88-2\rLeicester 72 88-9 Wessex 181 87-3\rUniversity College-Middlesex 115 87-8 Yorkshire 139 87-1\rGuy's-StThomas's 151 87 4 SouthEastThames 112 85-7\rManchester 225 84-9 Wales 129 84-5\rLeeds 92 84-8 SouthEastScotland 126 84-1\rQueen's, Belfast 241 84-6 Northem Scotland 42 83-3\rEdinburgh 155 83-9 South WestThames 127 82-7\rCharing Cross-Westminster 153 83-7 West Midlands 178 82-0\rTheLondon 101 82-2 NorthEastThames 115 81-7\rSt George's 39 82-1 Armed Forces 20 80-0\rSt Bartholomew's 104 81 7 North West 238 79 4\rSheffield 119 80-7 Mersey 84 72-6\rAberdeen 102 79-4 West of Scotland 271 70 5\rDundee 86 79-1 Scotland (Tayside) 50 68-0\rWales 95 76-8\rLiverpool 86 72-1 Not known 13 84-6\rGlasgow 218 69-3\rCambridge University\rSchool of Clinical\rMedicine, Addenbrooke's\rHospital, Cambridge\rCB22SP\rRichard Wakeford,\reducational adviser\rUniversity ofCambridge\rLocal Examinations\rSyndicate, Cambridge\rCBI 2EU\rJohn Foulkes, research officer\rSt Mary's Hospital Medical\rSchool, London W2 IPG\rChris McManus, senior\rlecturer in psychology as\rapplied to medicine\rMedical College ofSt\rBartholomew's Hospital,\rLondon EC1M 6BQ\rLesley Southgate, professor\rofgeneral practice\rmedical school and region of vocational training. To\ravoid bias, only first time United Kingdom and Irish\rbom takers, currently finishing their training, were\rincluded (n=3091).\rThe summary data are given in the table. Differences\rwere highly significant between medical\rschools (X2=92-46; df=26; p<0001) and training\rregions (x2=100-34; df=23; p<0001). Although\rmedical school and training region were related (42%\rof trainees carried out postgraduate training in the same\rregion as their undergraduate training), there are no\rdifferences overall between those who stayed in\rtheir region for postgraduate training and those who\rleft. Log-linear modelling using the program GLIM\r(generalised linear interactive modelling) showed that\rdifferences between training regions were significant\rafter differences between undergraduate schools were\rtaken into account (X2=56-6; df=23; p<0 001), and\rdifferences between undergraduate schools were significant\rafter differences between training regions were\rtaken into account (x2=49 7; df=26; p<001). The\rrank ordering of schools and regions was little altered\rby taking into consideration the effect of the other, and\rthe top and bottom six schools and regions remained\rthe same as in the tables.\rComment\rThe differences shown in the table may well be the\rresult of true differences in the training ability of\rmedical schools and regions. Altemative explanations\rare, however, not inconceivable. It might, for example,\rbe the case that good students from high scoring\rmedical schools and poor students from the low scoring\rschools consistently tend to opt for general practice.\rThe proportions of trainees from the different\rregions who attempt the examination vary (but\rthe available data on which to calculate these are\runsatisfactory).' The figures also take no account\rof the academic qualifications of students entering\rthe various medical schools and th