In an earlier issue of the New Zealand Medical Journal, we described the development of a clinical pathway for the investigation of patients with haematuria.1 The pathway used imaging and a urinary biomarker assay of bladder cancer (Cxbladder TriageTM (CxbT)) to identify patients who required a referral for specialist assessment to exclude the possibility of a bladder malignancy (refer to Appendix). The imaging modality requested depended on the type of haematuria and the age of the patient: renal ultrasound for microhaematuria and patients <40 and >85 years of age with macrohaematuria, and intravenous computed tomography (IVU-CT) for all other cases of macrohaematuria. The pathway did not include urine cytology in the initial laboratory screening tests and used the high negative predictive potential of CxbT combined with appropriate imaging to identify patients who did not require a secondary care appointment and flexible cystoscopy. Patients were only referred for urological assessment and cystoscopy if their CxbT index was positive and/or their imaging showed a bladder or other urinary-tract abnormality. Urine cytology could be requested at this stage at the discretion of the urologist. Patients with persistent microhaematuria, a urine total albumin:creatinine ratio >70, an eGFR <60 ml/min/1.73m2 or newly diagnosed hypertension were referred to a nephrologist for evaluation. The pathway was a departure from international clinical guidelines that recommend cystoscopy as the gold standard for diagnosis of bladder cancer.2,3 A previous retrospective analysis of a cohort of patients with haematuria showed that the clinical pathway would have detected all invasive urothelial carcinomas and that approximately one-third of patients could have safely avoided the need for an invasive flexible cystoscopy with negligible risk of a bladder cancer being missed.1

Since February 2018, patients with haematuria in the Canterbury region have been investigated under the guidelines of the new clinical pathway. To inform general practices of these changes, the pathway was made available on the Canterbury Community HealthPathways website.4 A clinical review of the effectiveness and safety of the new pathway was carried out prospectively during the first year of its introduction. This paper reports the findings of this review.

Patients and methods

The Health and Disability Ethics Committee, Ministry of Health, New Zealand, advised that the clinical review did not require ethical approval, as it constituted monitoring and improvement of usual patient care carried out by the Canterbury District Health Board (CDHB).

Patients

The clinical records of 889 patients who had a CxbT test for investigation of haematuria between 1 February 2018 and 31 January 2019 were reviewed. Five patients were excluded from the analysis: one with a bleed from a catheter, and four because of inadequate information on their medical records, two of whom had declined their hospital appointment.

Within the CDHB region there is a population of 567,870 (2018/19 projection). A single group of urologists provide urological services to all patients within the CDHB region, in both the public and private sectors. Thus, the patients described in this paper represent a nearly complete community capture.

Data collection

Data was collected on the remaining 884 patients in a non-blinded manner and included: demographic characteristics, type of haematuria, presence of risk factors for bladder cancer (smoking, previous history of bladder cancer and radiation therapy of the pelvis) and the findings of laboratory tests, imaging, cystoscopies, specialist assessment and histology. A cystoscopy and histology were required for diagnosis of bladder cancer. Macrohaematuria was defined as blood clearly visible in a midstream urine sample, and microhaematuria was defined as >20x106L red blood cells in two of three samples collected seven days apart.

A follow-up review (median: 21 months; range: 16–27 months) was then carried out of all the patients who had not been diagnosed with bladder cancer to ensure that no malignancy had been missed. The patients’ electronic records of inpatient admissions, outpatient clinics and histology results, and a search of the New Zealand Cancer Registry, were included in this review.

Finally, all the new diagnoses of bladder cancer in the CDHB region for a subsequent 15-month period from 1 February 2019 to 30 April 2020 were identified from surgical and pathology records. The patient records were reviewed to see whether any of these patients had previously been investigated by the new pathway for haematuria and had a bladder cancer missed.

Statistical analysis

The flow of patients through the three arms of the clinical pathway, grouped according to whether they had microhaematuria, macrohaematuria or were aged <40yr or >85, was examined graphically. The diagnostic accuracy of the various investigations in the pathway to detect bladder cancer was determined by calculation of accuracy, sensitivity, specificity and negative predictive value (NPV). The evaluation of CxbT used a segregation index cut-off value of <4.0 to indicate specialist assessment was required.5

Statistical power analysis showed that at least 600 patients would allow a precise estimate of the diagnostic accuracy of the indices, assuming a target sensitivity of 90% and a precision of sensitivity measurement of 10%, and giventhe prevalence of bladder cancer observed in the review patient cohort of 5.8%.

Results

The clinical and demographic characteristics of the 884 patients are summarised in Table 1. All the patients lived in the funded area of the Canterbury DHB and were predominantly middle-aged or older, with 66% being male. Two-thirds of the patients presented with macrohaematuria. Approximately 43% of the patients were classified as having an increased risk of developing bladder cancer because of their smoking history (41%), a previous history of bladder cancer (0.001%) or having previously received radiation therapy of the pelvis (2%).

Table 1: Demographic and clinical characteristics of the 884 patients.

The flow of the patients through the three arms of the clinical pathway and the diagnoses for the cause of haematuria are shown in Figure 1. One hundred and seventeen patients were still referred to a specialist urologist despite their CxbT being normal and their imaging indicating no bladder cancer was present. Seventy patients were diagnosed with another urological condition, while in 47 cases the cause of haematuria could not be identified.

The clinical pathway detected all but one case of bladder cancer, with 48 patients (macrohaematuria: n=44; microhaematuria: n=4) diagnosed with a histologically confirmed (n=46), or clinically possible (n=2), transitional cell carcinoma (TCC) and three with a small cell carcinoma. The stage and grade of the 48 TCCs were low grade (pTa n=20 and CIS n=1), and high grade (pTa n=14, pT1 n=8 and pT2 n=3). Two further malignancies were designated cT2-3, as neither had a cystoscopy or histology (one was frail and 91 years old, and the other had widespread metastatic bowel cancer). The 51 cases of bladder cancer represented a prevalence rate of 5.77%.

Three hundred and forty-eight (39%) patients with haematuria and normal CxbT and radiology were managed in primary care alone. Two hundred and eighty-nine patients had a sample collected for urine cytology, with 64% of these requests made by a urologist at the time of the hospital clinic visit.

Figure 1: Flow of the 884 patients through the three arms of the haematuria clinical pathway.

The ability of CxbT, imaging and the combination of CxbT and imaging to detect bladder cancer and the diagnostic accuracy of these investigations in the study cohort are summarised in Table 2. Five cancers were missed by CxbT. One of these was a superficial, high-grade lesion (pTaHG, 1cm in size). A second was considered from cytology to be high grade, but clinically it was thought to be most likely of bowel origin in a patient with extensive metastatic bowel cancer. No cystoscopy was performed, nor was a histology taken, in this patient. The other three cancers were low-grade pTa lesions. Although CxbT had relatively high sensitivity to detect a bladder cancer and a high NPV of 98.9% to exclude the possibly of a lesion, the test had a low specificity with 39% of the tests returning a false-positive result. This level of diagnostic accuracy for CxbT is similar to that reported previously.6–8 The mean of the CxbT segregation index was 3.97 (SD=0.92; range: 1.97–10.0), with 466 patients (53%) having a score less than the triage cut-off of 4.0. Approximately 10% of the patients (85 of 884) required a repeat CxbT assay because of quality control failures, mainly caused by interference of inflammatory products or a large number of white blood cells. Of these 85 patients, the repeat test provided a result in 28, and the second test did not meet quality standards in 15, while a repeat test was not requested in 42.

Imaging failed to detect 13 lesions (ultrasound: n=7; CT-IVU: n=5; ultrasound + CT-IVU: n=1), with nine of these lesions classified as low-grade lesions (pTa) and four as high-grade lesions (pT1: n=3; assessed by cytology: n=1). When combined, the pathway of CxbT and imaging had a sensitivity of 98.1% and an NPV of 99.9% and failed to detect a solitary, small, low-grade pTa bladder cancer.

Flexible cystoscopy detected bladder cancer in 48 patients, while the remaining three cancers were observed subsequently by rigid cystoscopy carried out at the time of transurethral resection of a bladder tumour (TURBT). One patient with widespread metastatic disease from bowel cancer did not have a cystoscopy. 

Table 2: Evaluation of the diagnostic accuracy of CxbT, imaging and cystoscopy to detect bladder cancer. The diagnostic parameters are expressed as percentages (95% confidence interval).

Table 3 summarises the causes of haematuria and the percentage of each cause detected by the clinical pathway. Renal stones, vascular prostate and inflammatory conditions were the most common causes of haematuria. Inflammatory renal disease was the cause of microhaematuria in six cases, all of whom were referred to a nephrologist for management.3 In approximately one-third of patients, the cause of haematuria was not determined.

Table 3: Causes of haematuria in the study cohort.

Follow-up review of medical records of the 833 patients who had not been diagnosed with bladder cancer showed that one patient had been treated for a small papillary lesion on the bladder neck six months following a nephroureterectomy for a high-grade TCC of the kidney. No other patient had been admitted to hospital for management of a new bladder malignancy. A search of the New Zealand Cancer Registry confirmed this finding.

A review of all the subsequent bladder cancers diagnosed in the following 15 months identified 111 patients with a first diagnosis of bladder cancer, confirmed histologically. Seventy-one percent were identified through GP workup of haematuria, 23% incidentally through radiology, 2% incidentally through cystoscopy and 4% through secondary care investigation of haematuria. Of those coming through the GP workup of haematuria, 42 (53%) had a successful CxbT test and eight had a “quality control” failure of the test. Two patients had a CxbT test that was <4 and a suspicious radiology. Both were found to have low-grade pTa malignancies. Only one patient had a bladder cancer missed in an earlier workup of haematuria. This patient had an ultrasound, but no CxbT, and a flexible cystoscopy that showed no abnormality. The cystoscopy was repeated three months later due to continued haematuria and the cancer was found.

Discussion

Successful treatment of invasive bladder cancer relies on early detection of the malignancy. Flexible cystoscopy supported by the findings of imaging and urine cytology have been the “gold standard” diagnostic procedures for investigation of patients with haematuria and a possible bladder cancer.2,3 However, the invasive nature of cystoscopy and its relatively poor cost performance,9 coupled with the low diagnostic accuracy of cytology for low-grade bladder cancers, emphasises the need for an improved clinical approach for investigation of haematuria.10 To this end, over the last 10 years a large number of urinary biomarkers have been developed and assessed for the diagnosis and monitoring of bladder cancer.11–14 The range and type of these biomarkers reflects the complex aetiology and development of bladder cancer, with assay profiles based on either the levels or presence of specific microRNAs,15,16 proteins,17 metabolites18 or extracellular vesicals.12 Despite extensive evaluation, none of these molecular biomarkers are included in clinical guidelines or recommended for use in daily clinical practice.16,19–22

The clinical pathway for investigation of haematuria evaluated in this paper was developed against this background and included the addition of CxbT in the initial laboratory investigations. This assay measures the level of five miRNAs to calculate a segregation index for stratification of patients.6,7 miRNAs are promising candidates for the diagnosis of bladder cancer as they are involved in several processes associated with the development of these malignancies, including proliferation, invasion, migration and apoptosis.23 Evaluation of the pathway during its development showed that CxbT improved the risk stratification of patients with haematuria and identified those in whom it was safe not to undertake cystoscopy.1 The prospective review of the pathway and follow-up of the patients described in this paper confirmed that the clinical pathway was an effective clinical algorithm for investigation of haematuria and that it reliably identified patients with bladder cancer. Interestingly, the 5.8% prevalence of bladder cancer in the current review was lower than that observed in the patient cohort used for development of the clinical pathway (9.2%).1 The reasons for this difference may be natural variation, an increased awareness of haematuria due to the communication around the change to the HealthPathway, or as a consequence of the relatively small size of the patient cohorts. We are currently carrying out reviews of the cost effectiveness, equity of access to haematuria assessment and compliance with the clinical pathway.

The risk of avoiding a cystoscopy is the possibility of missing a significant bladder cancer. Our results showed a false-negative CxbT result was obtained in five of the 52 patients diagnosed with bladder cancer, with one of these being a superficial (pTa) high-grade lesion, one likely to be a metastasis from a bowel cancer,not a transitional cell malignancy, and with the remaining three being low-grade pTa malignancies. Imaging detected the bladder cancer in four of these patients, while in the remaining patient a small, low-risk cancer (pTaLG) was detected three months later by cystoscopy following continued haematuria. These findings emphasise that an accurate risk stratification of patients with haematuria cannot be achieved using a urinary biomarker alone, and that the use of these tests in clinical pathways must be supported by appropriate imaging.

The adoption of the haematuria assessment pathway allowed 39% of patients to be fully cared for in primary practice, thus avoiding referral to secondary care. While this benefits the patient and the system, it is important to be certain that the clinical risk to the patient is not increased. Of the 884 haematuria patients undertaking the pathway (CxbT and radiology), only a single low-risk (low-grade pTa) cancer was missed. The NPV of 99.9% is consistent with previously published results.1,5 The follow-up review of the clinical records of the 832 patients who did not have a bladder cancer detected by the new pathway identified a single similar low-risk cancer in a follow-up cystoscopy of a patient in whom the pathway had previously diagnosed an upper tract transitional cell malignancy. In addition, a review of all the newly diagnosed bladder cancers in the CDHB region over the following 15 months identified only one bladder cancer missed during a previous haematuria investigation. The haematuria investigation in this case had not followed the “pathway”, and had not had a CxbT, but did have a cystoscopy that did not identify the cancer. Therefore, it is concluded that patients with haematuria and a negative “pathway” workup (CxbT and radiology) have a negligible clinical risk of missing a significant bladder cancer.

Finally, the argument could be made that there are other significant causes of haematuria. Of these, upper tract tumours and urinary tract stones are the most significant. These are readily identified on radiology as a part of the haematuria pathway. When evaluating the findings of the follow-up review of approximately 21 months, it is important to take into account that the development of non-invasive papillary carcinomas is relatively slow24 and may vary according to tumour grade and stage.24,25 It is therefore possible that a lesion may not have been identified within this time period.

In conclusion, we are not aware of any other study that has prospectively assessed the inclusion of a urinary biomarker to improve the risk stratification of patients with haematuria. The findings of our study add to the increasing evidence that biomarkers have a place in the assessment of haematuria, but that the results of these assays need to be supported by imaging of the bladder. Our study demonstrates that, when they are provided with CxbT results in combination with imaging, clinicians are able to reliably identify patients who can be assessed in primary care without the need for a secondary care referral and a flexible cystoscopy. We consider that in a well-integrated health system the management of these test-negative patients can be safely undertaken by general practitioners with support and oversight provided by a specialist urologist.

Appendix

Appendix Figure 1: Extract from Canterbury Community HealthPathways used by general practice for locally agreed guidance on over 800 conditions and situations. Each piece of underlined blue text represents a dropdown providing further clinical and process information.

‍