The accuracy of one-stop diagnosis for patients presenting to a symptomatic breast clinic

AMIN ELTAHIR*, JIBRIL A. JIBRIL*, JANET SQUAIR#, STEVEN D. HEYS*, ANTOINNE K. AH-SEE*, GILLIAN NEED-HAM†, FIONA J. GILBERT†, HEATHER E. DEANS†, MARY E. MCKEAN§, LOUISE M. SMART§ and OLEG EREMIN*
*Professorial Surgical Unit, #Departments of Public Health, †Radiology and §Cytology, University of Aberdeen and Aberdeen Royal Hospitals NHS Trust, Aberdeen, U.K.

To minimise delay in diagnosis and reduce patient anxiety, triple assessment with immediate reporting has been used in our symptomatic breast clinic since 1991. This article examines the accuracy of the diagnostic modalities used and the efficacy of the “one-stop” diagnostic policy. The data on 1 110 new patients presenting to the symptomatic breast clinic between January and July 1993, were analysed and subsequent three year follow-up and outcome established. Fine needle aspiration cytology (FNAC) gave the highest predictive value (97.3%) with a sensitivity of 93.5% and a specificity of 98.1%. Ultrasonography provided a 97.0% prediction with a sensitivity of 88.9% and a specificity of 97.4%, whereas mammography had a prediction of 96.4% with sensitivity of 93.2% and a specificity of 96.7%. When the mammogram or ultrasound scan were reported as unequivocally benign (M1), there were no missed cancers. The false positive and false negative rates for FNAC were 0% and 1.4%, respectively. Following assessment, a diagnosis was made in 96% of patients. Sixty-two percent of the patients were discharged at the first clinic visit. Four breast malignancies were subsequently diagnosed in the discharged group; two with new microcalcifications due to ductal carcinoma in situ, one with invasive disease in a different quadrant of the breast from that originally investigated, and in one patient the cancer was missed. Conclusion: A “one-stop” symptomatic breast clinic provides an accurate and effective means of establishing a correct diagnosis.

Keywords: breast clinic, breast disease, mammogram, breast ultrasound, cytology, triple assessment, one-stop diagnosis

Breast disease and its symptoms generates much media attention. This had been heightened by the implementation of a National Breast Screening Program, which has contributed to increasing public awareness of breast cancer as a major and common public health issue. Patients have a high expectation for the successful and efficient management of their symptoms. There is also an increasing professional requirement (eg, through published guidelines) on the part of clinicians for improved health care delivery for these patients.¹ This public and professional awareness and concern has led to a change in the referral pattern of patients with breast symptoms. General practitioners are no longer willing or indeed able to reassure patients without specialist assessment. An increasing number of patients, therefore, are being referred for specialist opinion with the benign to malignant ratio constantly rising.^2,3

The majority of patients referred to a breast clinic have benign disease.⁴ Most of these patients, however, are in a state of heightened anxiety until they have undergone specialist assessment, the necessary investigations and eventual reassurance.^5,6

To minimise delay and, therefore, reduce anxiety in the majority of cases presenting to hospital and to avoid unnecessary outpatient follow-up and open biopsy, many breast clinics have evolved a policy of triple assessment with immediate reporting to provide a “one-stop” diagnostic service.^2,7 Patients are evaluated by history/physical examination, imaging (mammography, breast ultrasonography) and fine needle aspiration cytology (FNAC) to establish a diagnosis and management plan for each patient on the day of the clinic visit. For this policy to be effective and safe, a continuous and vigilant mechanism of quality assurance is necessary to enable each centre to ensure appropriate utilisation of resources and that patients’ care is not compromised.

This article reports on the accuracy and efficacy of the “onestop” clinic policy in our hospital by matching the results of individual investigative modalities and the action taken at the initial clinic visit with the patients’ outcome three years later. The three year interval was chosen to allow detection of any cancers missed at the original clinic visit (ie, false negative “one-stop” evaluations).

The data on all new patients presenting to the symptomatic Breast Clinic at Aberdeen Royal Infirmary between 1st January and 31st July 1993 were retrospectively analysed and subsequent three year follow-up and outcome established. The clinic, at that time, was run once weekly and was staffed by three consultant breast surgeons, one senior registrar, one research registrar, a specialist nurse counsellor, one Sister and three staff nurses. Patients underwent one or more of the following investigations: mammography, breast ultrasonography and FNAC. A consultant radiologist and one or two consultant cytopathologists (depending on the workload and supported by a technician) provided immediate reporting of the diagnostic modalities.

All patients were examined by a consultant surgeon with a specialist interest in breast disease. Where the patients were initially examined by a trainee, a consultant surgeon also saw the patient to verify the clinical findings. The clinical examination was performed using standard examination techniques in the presence of a clinic nurse who acted as a chaperone. The clinical impression following examination and evaluation of the symptoms was documented as benign, indeterminate or malignant.

All patients who were 35 years old and above, and who did not have a mammogram within a year of their referral to the clinic, were routinely subjected to two-view mammography, unless they were pregnant or had severe mastalgia and could not tolerate the procedure. Mammography was performed prior to the clinic visit if the patient lived in Aberdeen, or on the day of the visit if resident outwith the city. The mammograms were performed using a Siemens Mammomat unit and Kodak Min-R film/screen combination and Kodak chemicals. Oblique and cranio-caudal views were obtained and supple-mented with additional views as required. The mammograms were reported by one of three consultant radiologists with a specialist interest in breast disease. Reports were assigned a score (M1-M5).⁸ The mammograms were reviewed at the clinic and, if necessary, discussed with the radiologist.

When indicated, an ultrasound scan was performed in the clinic by the radiologist, who provided an immediate report. The scan was performed with a Siemens Sonoline SL 250 unit with a 7.5 MHz linear array transducer with large footprint. Indications for ultrasonography were: the presence of a palpable abnormality in a woman less than 35 years of age, a non-contributory mammogram (eg, dense background pattern, no obvious abnormality at site of clinical concern), or when the mammogram was contraindicated (eg, pregnant patient). Targeted ultrasound examination was employed; only the breast tissue under clinical evaluation was examined. No attempt was made to examine the whole of each breast.

This was performed in all patients with a palpable abnormality such as a discrete lump, circumscribed area of thickening or asymmetrical nodularity. In addition, patients with suspected Paget’s disease underwent nipple scrapings. Cytology was also performed on certain nipple discharges (eg, blood stained, single duct). Two consultant cytopathologists with an interest in breast disease provided this service. The cytopathologists performed the aspirate and smears were stained using “Diff-Quik” rapid and Papanicolou methods. An immediate report was given using the standard reporting categories: acellular/inadequate, benign, atypical, suspicious and malignant (C1-C5).⁹

During the study period (January - July 1993), 1 218 patients were referred to the Clinic. Of these, 1 110 case notes were retrieved and available for assessment. By cross referencing our data with that stored for all cases of breast cancer by the Regional Office of the Scottish Cancer Therapy Network (SCTN), it was possible to establish that none of the 108 unavailable records had documented breast cancer, either at the clinic visit or subsequently. All the patients were female and the age range was 15 to 91 years.

The definitive diagnosis for each patient was based on the histopathological diagnosis in patients who subsequently underwent diagnostic biopsies or definitive surgery. For the majority of patients, however, diagnosis was based on the absence of further symptoms or a diagnosis of malignancy at the end of a three year follow-up period. The histopathological result or the clinical absence of malignancy within three years was matched to the original results obtained at the clinic by examination, mammography, ultrasonography and FNAC, to determine the accuracy of these tests.

Mammography was performed in 65.6% (728/1 110) of the patients. In 27.2% of these latter patients, the mammographic report was scored as indeterminate (score M3) and, therefore, unable to define the patient’s diagnosis more precisely. Ultrasonography was performed in 36.9% (410/1 110) of the patients with the report being scored as indeterminate in 11.2%. FNAC was performed in 34.5% (383/1 110) of the patients. This was acellular (using strict criteria) in 30% of the patients tested.

The sensitivity, specificity, and overall correct prediction of each modality of investigation were calculated and are shown in Table 1. The false positive and false negative rates for mammography, FNAC and ultrasonography are shown in Table 2. Ten percent (110) of the patients had all the investigate modalities performed; when all three tests suggested a malignant outcome, the concordance rate was 100%.

Test	Sensitivity (%)	Specificity (%)	Prediction (%)
Clinical examination (n=1 070)	88.7	99.1	98.5
Mammography (n=728)	93.2	96.7	96.4
Ultrasound (n=410)	88.9	97.4	96.8
FNA cytology (n=426)	93.5	98.1	96.9

Table 2: False negative and false positive rate of various assessment modalities

	False negative rate		False positive rate
Clinical Examination (n=1 070)	6/969	0.61%	9/56	10.7%
Mammography (n=728)	0/183	0.0%	1/24	04.0%
Ultrasound (n=410)	0/82	0.0%	1/11	09.1%
FNA Cytology (n=426)	3/215	1.4%	1/42	02.4%

Following initial assessment at the clinic, a diagnosis was made in 96% of the patients with 59 patients (5%) being diagnosed as having breast cancer. Sixty two percent of the patients were discharged following assessment. Figure 1 shows the reasons why the remaining 38% of patients were not discharged.

Figure 1: Management outcome for patients not discharged at first clinic (n=422)

During three years of follow-up four breast malignancies were diagnosed. Two were detected by a subsequent mammogram carried out by the local NHS Breast Screening Service. The latter revealed new microcalcification (absent on original clinic mammograms) which were subsequently found to be ductal carcinoma in situ. A third patient re-presented 2½ years later with invasive disease in a different quadrant of the breast from that originally investigated. In only one patient (85yrs old and frail) was the diagnosis of cancer apparently missed at the initial evaluation.

The most important feature of a “one-stop” diagnostic service is that a management plan can be determined and discussed with the patient at the initial outpatient attendance, thus, giving immediate relief and reassurance to the majority of patients with benign disease, and obviating the need for review appointments. Previous studies in our unit⁷ and elsewhere^2,10,11 have focused on the accuracy of triple assessment in dominant breast lumps that were subsequently excised for histological examination. These results have shown that when the results of triple assessment are evaluated concurrently, a reliable diagnosis can be established, thereby, avoiding unnecessary biopsy of benign lesions which can be managed conservatively.^2,7,10,12,13

Many patients now present to breast clinics without a discrete lump that is suitable for excision biopsy. These patients, like those with discrete lesions, require the same degree of specialist evaluation and assurance. The result of a diagnostic biopsy, therefore, can not be used as an end-point for determining the accuracy of the “one-stop” clinic policy. In our opinion, the only way that this can be effectively achieved, for all attendees, is by matching the long-term outcomes of these patients with the results of the investigations performed at the initial clinic visit.

Although clinical examination gave the highest prediction, its sensitivity (ability to detect disease when present) was the lowest, making it the least reliable in the context of a diagnostic clinic.

Mammography remains the most important imaging investigation for breast disease and is the standard against which newer imaging modalities are compared. Because of its general application and widespread use, there is a high proportion of indeterminate scores (M3) in most series. Our figure of 27.2% falls within the range reported from other centres.¹⁴

Ultrasonography is generally considered to have a limited role in the diagnosis of breast cancer.^15,16 Its application is quite different from that of mammography, offering targeted evaluation of a focal area rather than the whole of the breast. As a consequence of this different application, the percentage of indeterminate reports is lower than for mammography. Furthermore, mammograms (when performed) are available to the radiologist at the time of performing and reporting the ultrasound scan. Although the addition of ultrasound in a multivariate analysis did not improve the diagnostic yield for malignant cases (data not shown), its availability provided more qualitative information and improved the overall diagnostic accuracy and confidence with which the patients with benign disease were reassured.

FNAC of palpable lesions is an established diagnostic technique with a high degree of accuracy.^7,10,17,18 It provided the highest overall prediction of both benign and malignant dis-ease in our clinic. The high percentage of acellular specimens may be explained by our preferred pattern of application and strict definition of acellularity but, nonetheless, compares well with figures from other UK centres.^2,18,19

The accuracy of all the diagnostic modalities used at our clinic is one of the best reported in the literature^42,10,20 and has improved since last evaluated in our clinic in 1990.⁷ Of significance is the fact that, when the mammogram or ultrasound scan were reported to be unequivocally benign (ie, M1), there were no missed cancers. Similarly, when the FNAC report was categorically malignant (ie, C5), there were no false positive outcomes. Interestingly, the three instances of false negative FNAC occurred in patients with large (>40mm) discrete lumps that were considered malignant on clinical examination and imaging. This underscores the necessity for triple assessment as even FNAC has its limitations, especially when there is sampling error and when the cancer is histologically lobular in type.^21,22

A diagnosis was achieved in 96% of patients at the first clinic consultation. In the absence of comparable published studies from other centres it is difficult to evaluate the effectiveness of this performance. The high benign to malignant ratio is a reflection of the increasing anxiety in relation to breast symptoms and need for reassurance of both patients and their general practitioners.

In this study, 62% of the patients were discharged following assessment by the end of the “one-stop” clinic. There are currently no guidelines as to what the optimum discharge rate should be. Reports on percentages of discharge from other centres are all on a selected group of patients.10,20 Of the remaining 38% of patients who were not discharged by the end of the “one-stop” clinic, 30% were followed up because of our policy of reviewing all patients following cyst aspiration and those commenced on treatment for severe mastalgia. The clinical benefit and cost effectiveness of this policy is currently under review. Another 30% were reviewed for further patient reassurance. The discharge rate could be increased in this latter group as none of these patients were shown to have developed significant disease during the follow-up period. Qualitative measurement of the value of review to psychomorbidity in this group of patients would further define this policy.

The one patient whose diagnosis of breast cancer was missed was an 85-year old lady who presented with an indistinct area of nodularity in her left breast. Mammography was reported as probably benign (M2) and FNAC was acellular (C1). Ultrasound was not performed. She re-presented 2½ years later with a 2.5cm lump in the same area of the breast and this was found to be malignant (C5) on FNAC. It is difficult to determine (in a retrospective study) the extent to which the patient’s concurrent medical condition modified the initial assessment, as even when subsequently diagnosed she was considered too frail for surgery. She was treated with Tamoxifen and remains alive with local tumour control. In retrospect, this patient could have been followed-up with an ultrasound guided FNAC. To reduce the risk of sampling error, our current policy is to perform image-guided FNAC when dealing with discrete lumps that are acellular (C1) on free-hand FNA.

In conclusion, we believe that in our practice, the “one-stop” diagnostic service provides a reliable and accurate means of establishing a rapid diagnosis and is a safe and efficacious process for managing the ever increasing number of patients presenting with breast symptoms.

Correspondence: Professor O Eremin, Lincoln and Louth NHS Trust, County Hospital, Greetwell Road, Lincoln LN2 5QY, UK

ORIGINAL ARTICLES

The accuracy of “one-stop” diagnosis for 1 110 patients presenting to a symptomatic breast clinic