colorectal cancer pathology reporting: aregional auditon standards for pathology reporting...

3JClin Pathol 1997;50:138-142

Colorectal cancer pathology reporting: a regionalaudit

A D Bull, A H B Biffin, J Mella, A G Radcliffe, J D Stamatakis, R J C Steele, G T Williams

AbstractAims-To audit the information content ofpathology reports of colorectal cancerspecimens in one National Health Serviceregion.Methods-All reports of colorectal cancerresection specimens from the 17 NHS his-topathology laboratories in Wales during1993 were evaluated against: (a) standardspreviously agreed as desirable by patholo-gists in Wales; and (b) standards consid-ered to be the minimum required forinformed patient management.Results-1242 reports were audited. Therewas notable variation in the performanceof different laboratories and in the com-pleteness of reporting of individual itemsof information. While many items weregenerally well reported, only 51.5% (640/1242) of rectal cancer reports contained astatement on the completeness of excisionat the circumferential resection marginand only 30% (373/1242) of all reportsstated the number of involved lymphnodes. All of the previously agreed itemswere contained in only 11.3% (140/1242) ofreports on colonic tumours and 4.0%(40/1242) of reports on rectal tumours.Seventy eight per cent (969/1242) of co-lonic carcinoma reports and 46.6% (579/1242) of rectal carcinoma reports met theminimum standards.Conclusions-The informational contentof many routine pathology reports oncolorectal cancer resection specimens isinadequate for quality patient manage-ment, for ensuring a clinically effectivecancer service through audit, and for can-cer registration. Template proforma re-porting using nationally agreed standardsis recommended as a remedy for this,along with improved education, review oflaboratory practices in the light ofcurrentknowledge, and further motivation ofpathologists through their involvement inmultidisciplinary cancer managementteams.(J Clin Pathol 1997;50:138-142)

Keywords: colorectal cancer; histopathology; audit;quality; guidelines; protocols.

Histopathological reporting of resection speci-mens for colorectal cancer provides importantinformation both for the clinical managementof the affected patient and for the evaluation ofhealth care as a whole. For the individualpatient, it confirms the diagnosis and describes

variables that affect prognosis,' notably theextent of the disease (the pathological stage)and the completeness of local excision, both ofwhich inform future clinical management. Forexample, a number of randomised trials havedemonstrated the benefit of adjuvant chemo-therapy for tumours that have spread to theregional lymph nodes (Dukes' C cases).2 Inrectal cancer, there is evidence that involve-ment of the circumferential resection margin isan important predictor of local recurrence andpatient survival3 which might identify patientswho would benefit from postoperative adjuvanttherapy.4 In addition, pathology reports areoften used by insurance companies in assessingfinancial risk when patients seek insuranceafter a diagnosis of cancer. For health careevaluation, pathology reports provide poten-tially robust information for cancer regis-tration, for clinical audit, for assessing theaccuracy of new diagnostic and preoperativestaging techniques, and for ensuring compara-bility of patient groups in clinical trials. It iscrucial, therefore, that pathology reports ofcolorectal cancer specimens contain the infor-mation that is necessary to fulfil these func-tions, and that this information is accurate andcomplete.

Guidelines on the information content ofpathology reports in colorectal cancer havebeen published in standard textbooks ofpathology5 6 and by various expert workinggroups.7'10 The recommendations have variedin detail and complexity, ranging from compre-hensive lists of data items that encompass allvariables that could be of conceivable relevanceto prognosis to more focused, pragmatic 'mini-mum requirements' whose relevance to currentpatient management attempts to be moreevidence-based. However, there is no pub-lished information on the quality of colorectalcancer reporting in current histopathologypractice in the UK. In this paper, we presentthe findings of an audit of pathology reports ofcolorectal cancer specimens in all of the NHShistopathology laboratories in one UK region(Wales) during 1993, a study that was under-taken as part of a wide ranging audit of allaspects of colorectal cancer management in thePrincipality and in the Trent region of Englandwhose initial findings have been presentedrecently. "

MethodsThe population based audit of colorectalcancer management in Wales was directed byan expert working party set up by the RoyalCollege of Surgeons of England; the methods

Department ofPathology, UniversityofWales College ofMedicine, CardiffA D BullG T Williams

Clinical EffectivenessSupport Unit (Wales),PenarthA H B Biffin

Department ofSurgery, University ofNottingham,NottinghamJ MellaR J C Steele

Department ofSurgery, LlandoughHospital andCommunity NHSTrust, PenarthA G Radcliffe

Department ofSurgery, Bridgend andDistrict NHS Trust,BridgendJ D Stamatakis

Correspondence to:Professor G T Williams,Department of Pathology,University of Wales Collegeof Medicine, Heath Park,Cardiff CF4 4XN.

Accepted for publication5 November 1996

138

on July 27, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.50.2.138 on 1 F

ebruary 1997. Dow

nloaded from


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from

http://jcp.bmj.com/

http://jcp.bmj.com/

http://jcp.bmj.com/

Colorectal cancer pathology reporting: a regional audit

Table 1 Data items abstractedfrom pathology reports andusedfor audit

Colonic carcinomasLength of the specimenSize of the tumour (maximum dimension)Distance of tumour from nearest resection endMacroscopic description of the appearance of the tumourHistological type of the tumourHistological grade of the tumourExtent of tumour invasion into and beyond the bowel wallInvolvement of the resection ends by tumourInvolvement of regional lymph nodes by tumourNumber of lymph nodes involvedDukes' stage

Rectal carcinomasAll of the above plus:

Involvement of the circumferential excision planeMinimum distance of clearance at the circumferential

excision plane

Table 2 Minimum criteria for an adequate report

Colonic carcinomasHistological type of the tumourHistological grade of the tumourExtent of tumour invasion into and beyond the bowel wallInvolvement of the resection ends by tumourInvolvement of regional lymph nodes by tumour

Rectal carcinomasAll of the above plus:

Involvement of the circumferential excision plane

used will be described in detail elsewhere.Briefly, all patients with colorectal cancertreated in the 17 NHS district general hospitalsin Wales during the calendar year of 1993 wereidentified from a comprehensive search of anumber of data recording systems and, afterobtaining permission from the surgeons andhistopathologists involved, copies ofthe pathol-ogy reports were obtained by data collectors.The presence or absence of a statement onitems of information in the pathology reports(11 for colonic cancers, 13 for rectal cancers)was recorded on a proforma by a single surgicalresearch fellow GM), and transferred to a com-puter database by an optical mark scanner tobe analysed using the Statistical Package forSocial Sciences (SPSS for Windows, SPSSInc., Chicago, Illinois, USA). Data recordingwas validated by checking the abstracted infor-mation from a randomly selected 10% ofreports against the original pathology report byone of us (ADB).

0°. 100a)

o 80a).0E 60

z40

20

0

The items of pathology information re-corded are shown in table 1. They were basedon standards for pathology reporting of colo-rectal cancer that had been agreed previouslyby histopathologists in Wales in March 1992after discussion of published guidelines fromthe literature at regional pathology audit meet-ings. Pathologists from 15 of the 17 Welshlaboratories had attended at least one of thetwo meetings at which the standards were set,and all histopathologists in Wales were subse-quently circulated with the agreed conclusions.The informational content of all of the

pathology reports was audited against theagreed standards, both for Wales as a wholeand for each of the 17 laboratories. Thepercentage of reports containing a statementon each of the data items listed in table 1 wasobtained, and also the percentage containingstatements on all of the data items (11 forcolonic tumours, 13 for rectal tumours).Finally, the percentage of reports containingstatements on all of the data items in table 2was obtained, these less rigid criteria beingconsidered by us to be the minimum necessaryfor an adequate report. During the study, theidentity of the 17 laboratories was known onlyto one of the investigators (AHBB, an inde-pendent data manager) who, after the audit wascompleted, provided each of the laboratorieswith a report of its own performance, togetherwith that of each of the other laboratories(anonymised) and of the aggregated Welshlaboratories as a whole.

ResultsIn total, 1242 pathology reports were availablefor assessment, 57.2% ofwhich were resectionsfor colonic cancer and 42.8% for rectal cancer.The reports were issued from 17 NHS labora-tories staffed by 36 consultant histopatholo-gists. Although pathologists professing a spe-cial interest in gastrointestinal disease workedat some laboratories, the responsibility forreporting colorectal cancer specimens wasalways shared between all department consult-ants. Figure 1 shows the distribution of thenumber of reports obtained from each labora-tory, which varied from 27 to 152 (median 72).

Validation of the surgical research fellow'sabstracted data by a pathologist in a randomsample of 10% of the reports showed agree-ment with the interpretation of the originalreport for the vast majority of items ofinformation recorded. Virtually all of thediscrepancies related to statements which werenot completely clear in the original reports; insuch cases, the surgical research fellow hadnearly always interpreted the report generouslyand had recorded the data item as beingpresent.

Table 3 shows the percentage of reports con-taining statements on each of the individualitems of information defined by the standardsin table 1. It can be seen that performance var-ied considerably for different data items andbetween laboratories. All reports described thehistological type and grade of the tumour andalthough almost all stated the extent ofinvasioninto or beyond the bowel wall, this information

Laboratory

Figure 1 Histogram showing the number ofpathology reports obtainedfrom each of the17 laboratories in Wales.

139


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from

http://jcp.bmj.com/

Bull, Biffin, Mella, Radcliffe, Stamatakis, Steele, Williams

Table 3 Percentage of reports containing statements on individual data items audited

Worst performing Best performing Median for 17 Value for Wales asData item laboratory laboratory laboratories a whole

For all tumours:Length of specimen 84.1 100.0 98.4 96.7Tumour size 75.0 98.7 94.4 92.5Distance from resection end 52.5 82.9 75.2 72.9Appearance of tumour 36.1 98.4 92.0 84.9Histological type 100.0 100.0 100.0 100.0Histological grade 100.0 100.0 100.0 100.0Extent of invasion 77.8 100.0 98.6 96.7Resection end involvement 55.1 100.0 92.3 91.0Whether nodes involved 85.0 100.0 95.3 95.2Number of nodes involved 14.4 48.9 27.5 30.0Dukes' stage 12.5 93.8 73.6 74.9

For rectal tumours only:Circumferential plane involvement 8.3 78.1 57.6 51.6Measured circumferential plane clearance 0 28.1 7.7 12.1

Table 4 Percentage of reports fulfilling standards

Originally agreed standards (table 1) Mini

Laboratory Colonic tumours* Rectal tumours* Colon

A 2.6 (4) 0 (1) 44.7B 7.1 (6) 0 (1) 50.0C 0 (1) 0 (1) 52.0D 9.4 (9) 0 (1) 85.0E 28.3 (17) 0 (1) 93.5F 5.1 (5) 8.6 (15) 82.1G 10.0 (11) 2.9 (8) 68.6H 0 (1) 5.9 (12) 95.8I 14.9 (13) 8.8 (16) 54.5I 0 (1) 0 (1) 80.0K 13.3 (12) 3.0 (9) 88.0L 16.2 (14) 4.7 (11) 74.5M 7.4 (7) 8.3 (14) 70.4N 9.8 (10) 3.2 (10) 87.5O 8.8 (8) 0 (1) 86.5P 19.4 (15) 7.1 (13) 88.9Q 19.7 (16) 10.1 (17) 92.9All Wales 11.3 4.0 78.1

*Figures in brackets represent performance ranking of laboratories:

75 r-

U,

0

E

0)

50K

25

0

0

0

0

u440 60Colon tL

Figure 2 Regression plot of lalof cases fulfilling minimum stancarcinomas versus rectal carcinccorrelation (Spearman's rank c

p = +0.527,p <0.05).

was invariably present inlaboratories. Whether o0were involved was also g(95.2% overall) but foulgive this very importantevery case. However, thlymph nodes involvedbeing given in only 30% 4laboratory achieving a

Formal Dukes' staging o

in 74.9% of cases, varying notably betweenlaboratories from 12.5 to 93.8%.

mum standards (table 2) A statement on completeness of excision at

tic tumours*Rectal tumours* the ends of the specimen was given in 91 % of

(1) 5.9 (1) reports, but only two laboratories achieved(2) 8.3 (2) 100%. The distance from the tumour to the(3) 20.0 (3) nearest resection end was less frequently men-(10) 28.8 (4) tioned. However, completeness of excision at(9) 40.0 (6) the circumferential excision plane in rectal car-(5) 41.0 (7) cinomas was poorly reported, being mentioned(47) 47.1 (9) in only one half of all reports and the distance(8) 50.0 (10) of clearance being measured in less than one in(13) 51.5 (11) eight. Whereas the best performing laboratory(6) 58.3 (13) gave these two data items in 78.1% and 28.1%(12) 64.5 (14) of reports, respectively, seven laboratories(11) 65.2 (15) failed to describe the circumferential plane in(14) 71.4 (16)(15) 73.7 (17) >50% and four never mentioned the distance

46.6 of clearance.1 = worst; 17 = best. Table 4 shows the percentage of pathology

reports containing statements on all of the dataitems shown in tables 1 and 2 for each of the

° laboratories and for Wales as a whole. Theproportion of cases containing information on

0 o all of the items in table 1 represent those fulfill-0 ing all of the requirements originally agreed by

0 the pathologists in Wales before the audit com-0 0 menced. The results are disappointing with

° only 11.3% of colonic cancer reports and 4.0%0 of rectal cancer reports meeting all of the

0 agreed standards. Reports from two laborato-ries never contained all of the required dataitems, and no laboratory fulfilled the standardsin >20% of colonic tumours and > 10% of rec-tal tumours.

I The proportion of cases containing state-

80 100 ments on all of the items in table 2 representumours ( those fulfilling criteria considered by us to be

boratories'performance (o the minimum for an adequate report. Thisidards) for reporting colonic standard was only achieved in 78.1% of colonicomas. There is a positive cancer reports and 46.6% of rectal cancerorrelation coeffcient reports. No laboratory achieved 100% for

either tumour, and only six of the 17 achievedboth >80% for colonic tumours and >50% for

the reports of only five rectal tumours. There was no significant corre-r not the lymph nodes lation between laboratories' performancerenerally well reported against the agreed guidelines and their per-r laboratories failed to formance against the minimum standardsitem of information in (Spearman's rank correlation). For example,Le absolute number of laboratory J failed to satisfy the agreed criteriawas poorly reported, in any report but satisfied the minimum criteriaof cases overall with no in 80% of colonic cancer reports and 50% ofcompliance of >50%. rectal cancer reports. However, a significantf the tumour was given correlation (p < 0.05) was found between

n)

140


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from

http://jcp.bmj.com/

Colorectal cancer pathology reporting: a regional audit

laboratories' performance in reporting colonicand rectal cancers when the minimum stand-ards only were assessed (fig 2, Spearman's rankcorrelation).

DiscussionQuality pathology reports of colorectal cancerspecimens are central to ensuring clinicaleffectiveness in the management of this com-mon malignancy. Inadequate histopathologycan lead to inappropriate therapy and mislead-ing information for patients. It is, therefore, ofconsiderable concern that this audit of pathol-ogy reporting in Wales has highlighted impor-tant deficiencies in about one fifth of reports oncolonic cancer resections and in half of reportson rectal cancer resections. There is no reasonto believe that the situation in Wales is any dif-ferent from other health regions in the UK.The wide ranging clinical audit of colorectalcancer from which this study emerged alsoevaluated performance in the Trent region ofEngland." Although pathology data wererecorded for Trent, they are not available in aform that is strictly comparable with thosefrom Wales but, nevertheless, a preliminaryevaluation indicates that overall performancewas not significantly different (unpublishedobservations). Therefore, it is appropriate toconsider the nature of the Welsh audit in somedetail.

It is very important to make clear that weaudited only the information content of issuedreports. We have not investigated diagnosticprecision, the way that the specimens havebeen handled, the sampling of the specimensfor microscopic examination, or the accuracyof any of the data. Although all of these factorswill obviously influence the quality of theinformation that is presented in a pathologyreport, we have assessed only the completenessof the final report with regard to a definednumber of items of information.One strength of this audit is that it has

encompassed all of the pathology reports thatcould be retrieved in one year from all 17 labo-ratories in the region, and should therefore berelatively free from selection bias. Another fea-ture is that the information content of thepathology reports was abstracted for the auditprocess by an independent surgical researchfellow. This meant that it was interpreted by anunbiased end-user of the report rather than bya pathologist who might be tempted to inferinformation when it was not clearly presented.In fact, validation of the abstracted data in arandom sample of 10% by a pathologist identi-fied few discrepancies, and most of these indi-cated a generous interpretation of equivocalitems of information by the surgeon.The findings of this audit suggest that the

informational content of colorectal cancerreports issued by NHS pathology laboratoriesleaves much to be desired. Although it could beargued that some of the data items required arenot essential for informed patient manage-ment, such as the length of the specimen, thesize and appearance of the tumour, and thedistance of the tumour from the nearest resec-tion end, these are all items that would

generally be regarded as reflecting carefulexamination of the specimen and wouldcontribute to the value of the report in clinicalauditing of preoperative patient assessmentand surgical technique. Similarly, a statementof the pathological Dukes' stage might not beessential if the information required to derivethis is contained within the report, but someconcluding statement on pathological stage willgreatly facilitate cancer registration as registriesmove towards recording this routinely. How-ever, the most important deficiencies relate tothe poor description of circumferential resec-tion plane involvement in rectal carcinomasand the number of lymph nodes involved bymetastases. There is now strong evidence thatcircumferential margin involvement in rectalcancer is an important prognostic indicator,having high predictive value for both survivaland local recurrence.3 12 Accurate reporting ofcircumferential margin involvement is likely tohave considerable influence on the decisionwhether to use adjuvant radiotherapy orchemotherapy. Our finding that this infor-mation was given in only a half of rectal cancerreports suggests that effective patient treatmentmay be being compromised by incompletepathological reporting. Involvement of lymphnodes is another factor that is used to selectpatients for postoperative chemotherapy.2Moreover, the actual number of lymph nodesinvolved also has independent prognosticsignificance,' sufficient to warrant separation ofcases with less than four positive nodes fromother node-positive cases in both the TNM andthe Jass staging systems.'3 '4 It may be that thereluctance ofpathologists to report the numberof positive nodes in this audit is a reflection ofthe fact that this item does not contribute tothe Dukes' classification, which is most widelyused in the UK.The frequent failure of reports to contain

information on circumferential rectal margininvolvement and the number of positive lymphnodes, coupled with the lack of a stated Dukes'stage and a comment on the distance from thetumour to the resected end of the specimen inabout one quarter, are largely responsible forthe fact that few reports met all of thestandards originally agreed by the pathologistsin Wales. Because of this, the reports were alsoaudited against a minimum set of standardsthat were regarded as essential for postopera-tive management of the patient. Not surpris-ingly, this led to considerably improved results.Nevertheless, only 78.1% of colonic cancerreports and 46.6% ofrectal cancer reports werecomplete. The difference between these twofigures can be accounted for mainly by the poorreporting of circumferential margin involve-ment in rectal tumours, and this is clearly themost important factor requiring urgent reme-dial attention. Figure 2 shows that there was acorrelation between individual laboratories'performance in reporting colonic and rectaltumours, indicating that there are somelaboratories whose overall performance isclearly inferior to others. These differencesmay be related to variations in laboratoryworkloads and resources, but it is possible that

141


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from

http://jcp.bmj.com/

Bull, Biffin, Mella, Radcliffe, Stamatakis, Steele, Williams

the motivation of pathologists and their under-standing of the importance of some of theitems requiring description in the reports arealso partly to blame. For example, appreciationof the prognostic value of circumferential mar-gin involvement in rectal cancer resections hasonly been recognised relatively recently,5 16 andthe fact that the two best performing laborato-ries for reporting colonic cancer ranked fifthand eighth worst for reporting rectal cancer canbe directly attributed to their low frequency ofdescribing these margins.Although 21.9% of reports of colonic cancer

resections failed to fulfil all of the minimumcriteria, table 3 shows that each of the five indi-vidual items of information required to meetthis standard were generally well reported, withmedian values of 100%, 100%, 98.6%, 95.3%,and 92.3%, respectively. Apparently inadvert-ent omission of just one of the five items bypathologists who were otherwise diligent inrecording the information seemed to beresponsible for the failure of most of thereports. We believe that this is strong evidencefor a need for a formal reporting process usingstructured template proformas which, by serv-ing as a prompt for the pathologist, wouldensure that a statement on each critical item ofinformation was contained within the report. Anumber of such template proformas have beenpublished68 9 but informal discussions withpathologist colleagues suggests a reluctance touse them because they are perceived as beingtoo detailed for routine use, including items ofinformation of allegedly unconfirmed im-portance. Mindful of this, the histopathologistsadvising the Royal College of Surgeons ofEngland/Association of Coloproctology onguidelines for the management of colorectalcancer, after wide consultation with othergroups, have produced a minimum datatemplate proforma for reporting resectionspecimens."0 We advocate its use to alllaboratories, in the hope that it might become anational standard for minimum reporting thatwill be subject to regular review. We are notsuggesting that proforma reporting shouldentirely replace the current practice of usingfree text in reports. Pathologists must be free toinclude whatever information they regard asuseful and there will always be reports thatrequire description of special or unusualfeatures in specimens. However, we believe thattemplate proformas will go a long way towardsimproving the quality and completeness of theinformation provided for patient management.It will also facilitate recording of data on com-puterised information systems and its retrievalfor pathological input into audit, clinical trialsand cancer registration.Apart from the introduction of template

proforma reporting, what other mechanismsmight help to improve the current unsatisfac-tory position? We believe that continuingmedical education can contribute by informingpathologists of relatively new approaches todissecting and sampling resection specimens toobtain the maximum amount of information inan efficient way, 1 and in highlighting howclinical decisions on adjuvant therapy depend

more than ever on careful pathological stagingand evaluation of resection margins. Assess-ment of competence and understanding inthese areas should also be seen to be an impor-tant element of the accreditation process fortrainee pathologists. Managers should ensurethat laboratories have the resources to examineand report specimens to high standards and tomaintain these standards through internalaudit'7 -mechanisms already exist for encour-aging this through national laboratory accredi-tation. Finally, pathologists must be motivatedby being continually informed and reminded ofthe importance of their work and its quality bytheir surgeon and oncologist colleagues whoshould ensure that pathologists are seen as keymembers of the multidisciplinary teams whichwill deliver effective cancer care for the nextmillenium.' The recognition of individual con-sultants as lead pathologists for site-specificcancers within the developing cancer units andcentres should help considerably in this respect.

We wish to thank all of the consultant pathologists and staff ofthe 17 histopathology laboratories in Wales for providing accessto the pathology reports, and Rita Carter and Anne Whitlowwho collected the data from individual hospitals. We also thankthe Clinical Effectiveness Support Unit (Wales) for dataprocessing facilities and the Welsh Office and Department ofHealth for financial support.

1 Hermanek P, Sobin LH. Colorectal carcinoma. In: Her-manek P, Gospodarowicz MK, Henson DE, Hutter RVP,Sobin LH, eds. Prognostic factors in cancer. Berlin: Springer,1995:64-79.

2 Slevin ML. Adjuvant treatment for colorectal cancer: nomore room for nihilism. BMJ 1996;312:392-3.

3 Adam IJ, Mohamdee MO, Martin IG, Scott N, Finan PJ,Johnston D, et al. Role of circumferential margin involve-ment in the local recurrence of rectal cancer. Lancet 1994;344:707-11.

4 Krook JE, Moertel CG, Gunderson LL, Wieand HS,Collins RT, Beart RW, et al. Effective surgical adjuvanttherapy for high-risk rectal carcinoma. N Engl _'Med 1991;324:709-15.

5 Morson BC, Dawson IMP, Day DW, Jass JR, Price AB, Wil-liams GT. Gastrointestinal pathology. 3rd edn. Oxford:Blackwell Scientific Publications, 1990.

6 Rosai J. Ackerman's Surgical pathology. 8th edn. St Louis:Mosby, 1996.

7 Williams NS, Jass JR, Dixon MF, Quirke P, Johnson RJ,Robinson PJ, et al. Handbook for the clinicopathologicalassessment and staging of colorectal cancer. London: UKCo-ordinating Committee on Cancer Research, 1989.

8 Henson DE, Hutter RVP, Sobin LH, Bowman HE. Protocolfor the examination of specimens removed from patientswith colorectal carcinoma: a basis for checklists. Arch PatholLab Med 1994;118:122-5.

9 Association of Directors of Anatomic and Surgical Pathol-ogy. Recommendations for the reporting of resected largeintestinal carcinomas. Hum Pathol 1996;27:5-8.

10 Royal College of Surgeons of England/Association of Colo-proctology. Guidelinesfor the management of colorectal cancer.London: Royal College of Pathologists of England, 1996.

11 Stamatakis JD. An independent audit of colerectal cancer intwo UK health regions [abstract]. Br Jf Cancer 1996;74(Suppl XXVIH):50.

12 Ng IOL, Luk ISC, Yuen ST, Lau PWK, Pritchett CJ, Ng M,et al. Surgical clearance in resected rectal carcinomas; amultivariate analysis of clinicopathologic features. Cancer1993;71: 1972-6.

13 Hermanek P, Sobin LH (eds). TNM classification ofmalignant tumors. 4th edn. Berlin: Springer Verlag, 1992.

14 Stamatakis JD. An independent audit of colorectal cancer intwo UK health regions. [abstract]. Br JI Cancer 1996;74(Suppl xcviii):5.

15 Chan KW, Boey J, Wong SKC. A method for reportingradial invasion and surgical clearance of rectal cancer. His-topathology 1985;9: 1319-27.

16 Quirke P, Durdey P, Dixon MF, Williams NS. Local recur-rence of rectal adenocarcinoma due to inadequate surgicalresection: histopathological study of lateral tumour spreadand surgical excision. Lancet 1986;ii:996-9.

17 Campbell F, Griffiths DFR. Quantitative audit of thecontents of histopathology reports. JT Clin Pathot 1994;47:360-1.

18 Department of Health. A policy framework for commisioningcancer services. London: HMSO, 1994.

142


http://jcp.bmj.com

/J C


ebruary 1997. Dow

nloaded from

http://jcp.bmj.com/

Correspondence, Book reviews, Notices, Corrections

Colorectal cancer pathology reporting: a regionalaudit. A D Bull, A HB Biffin,J7 Mella, A GRadcliffe, D Stamatakis, R C Steele, G TWilliams. J Clin Pathol 1997;50:138-142.There were several errors in the abstract tothis article. The correct abstract is publishedbelow.

AbstractAims-To audit the information contentof pathology reports of colorectal cancer

specimens in one National Health Serv-ice region.Methods-All reports of colorectal can-

cer resection specimens from the 17NHS histopathology laboratories inWales during 1993 were evaluatedagainst: (a) standards previously agreedas desirable by pathologists in Wales;

and (b) standards considered to be theminimum required for informed patientmanagement.Results-1242 reports were audited.There was notable variation in theperformance of different laboratoriesand in the completeness of reporting ofindividual items of information. Whilemany items were generally well re-

ported, only 51.5% of rectal cancer

reports contained a statement on thecompleteness of excision at the circum-ferential resection margin and only 30%of all reports stated the number ofinvolved lymph nodes. All of the previ-ously agreed items were contained inonly 11.3% ofreports on colonic tumoursand 4.0% of reports on rectal tumours.Seventy eight per cent of colonic carci-noma reports and 46.6% of rectal carci-noma reports met the minimumstandards.

Conclusions-The informational con-

tent of many routine pathology reportson colorectal cancer resection specimensis inadequate for quality patient man-

agement, for ensuring a clinically effec-tive cancer service through audit, andfor cancer registration. Template pro-

forma reporting using nationally agreedstandards is recommended as a remedyfor this, along with improved education,review of laboratory practices in thelight of current knowledge, and furthermotivation of pathologists through theirinvolvement in multidisciplinary cancer

management teams.

In the same article reference 14 should haveread: Jass JR, Love SB, Northover JMA. Anew prognostic classification of rectal cancer.

Lancet 1987;i:1303-6, and not as published.

358

colorectal cancer pathology reporting: aregional auditon standards for pathology reporting...

Documents