Advice

Appendix

Appendix

Contents

Table 1: Overview of the Bishoff et al. (2014) study

Table 2: Overview of the Cooperberg et al. (2013) study

Table 3: Overview of the Cuzick et al. (2011) study

Table 4: Overview of the Cuzick et al. (2012) study

Table 5: Overview of the Cuzick et al. (2015) study

Table 6: Overview of the Freedland et al. (2013) study

Table 7: Overview of the Crawford et al. (2014) study

Table 8: Overview of the Shore et al. (2016) study

Table 9: Overview of the Warf et al. (2015) study

Table 10: Summary of the economic abstracts

Table 1 Overview of the Bishoff et al. (2014) study

Study component

Description

Objectives/hypotheses

To evaluate the prognostic utility of the CPP score derived from biopsy specimens in men treated with radical prostatectomy.

Study design

Retrospective cohort study.

Setting

3 cohorts: Martini Clinic (Germany; 2005–2006), Durham Veterans Affairs (USA; 1994–2005) and Intermountain Healthcare (US; 1997–2004).

Inclusion/exclusion criteria

Inclusion:

  • Men treated with radical prostatectomy.

  • Patients diagnosed with prostate adenocarcinoma without evidence of lymph node or bone metastases.

  • Formalin fixed, paraffin embedded tumour blocks containing a simulated (Martini Clinic) or diagnostic (Durham Veterans Affairs and Intermountain Healthcare) biopsy analysed at Myriad Genetics.

Exclusion:

  • Patients with preoperative PSA >100 ng/ml.

  • Patients with evidence of systemic disease or insufficient remaining tumour to generate a CCP score.

  • Patients who received neoadjuvant hormone therapy or radiation preoperatively.

Primary outcomes

Time to biochemical recurrence or metastatic disease.

Statistical methods

Survival analysis was performed with Cox proportional hazard methods using date of surgery as the starting time and time to BCR or metastatic progression as endpoints for the 3 cohorts combined. Effect size was measured by HR per unit of CCP score or another variable of interest with the 95% CI.

Patients included

582 patients total:

  • Martini Clinic: n=283; median age at surgery=63 years; 44% Gleason score ≥7; 77% clinical stage T1

  • Veterans Affairs: n=176; median age at surgery=62 years; 43% Gleason score ≥7; 62% clinical stage T1

  • Intermountain Healthcare: n=123; median age at surgery=62 years; 37% Gleason score ≥7; 58% clinical stage T2

Results

Median CCP score:

  • Martini Clinic: –0.4 (IQR –0.9, 0.2)

  • Veterans Affairs: 0.0 (IQR –0.4, 0.6)

  • Intermountain Healthcare: 0.3 (IQR –0.3, 0.9)

Combined analysis of all cohorts (total 582 patients) showed that CCP score was a strong predictor of biochemical recurrence on univariate analysis (HR per score unit 1.60, 95% CI 1.35 to 1.90, p=2.4×10−7) and multivariate analysis (HR per score unit 1.47, 95% CI 1.23–1.76, p=4.7×10−5). The combined cohort included 12 men with metastatic prostate cancer. Univariate analysis found that the score was predictive of metastatic disease (HR 3.35, 95% CI 2.89 to 9.92, p=2.1×10−8).

Conclusions

Increased CCP score derived from biopsy samples was associated with an increased risk in BCR in all 3 cohorts. CCP was also predictive of metastatic disease in univariate and multivariate analysis.

Abbreviations: BCR, biochemical recurrence; CCP, cell cycle progression; CI, confidence interval; HR, hazard ratio; IQR, interquartile range; PSA, prostate specific antigen.

Table 2 Overview of the Cooperberg et al. (2013) study

Study component

Description

Objectives/hypotheses

To validate the CCP score in predicting RP outcomes.

Study design

Prospective specimen collection, retrospective blinded evaluation design for biomarker validation.

Setting

USA, 1994–2011.

Inclusion/exclusion criteria

Inclusion:

  • Patients who underwent RP without neoadjuvant or adjuvant therapy.

  • Patients with at least 5 years follow-up after RP.

Exclusion:

  • Patients diagnosed prior to 1994.

Primary outcomes

The value of the CCP score.

The clinical utility of the CCP score.

Statistical methods

Association between the CAPRA-S score and the CCP score was examined using scatter plots and Pearson's correlation. Kaplan–Meier survival analysis was performed and multivariable Cox regression was used to assess the utility of the score.

Patients included

n=413; median age 59 years, IQR 54-63; 58% with Gleason score ≥7

Results

82/413 (19.9%) experienced recurrence.

The hazard ratio for each unit increase in CCP score was 2.1 (95% CI 1.6 to 2.9, p<0.001). Hazard ratio was 1.7 (95% CI 1.3 to 2.4, p<0.001) after adjustment by CAPRA-S score.

Conclusions

The CCP score was predictive of BCR regardless of the clinical risk group. The CCP score was weakly but significantly correlated to the CAPRA-S score (r=0.21, p<0.001). The combination of the 2 scores was more predictive than the CAPRA-S score alone.

Abbreviations: BCR, biochemical recurrence; CAPRA-S, Cancer of the Prostate Risk Assessment post-Surgical; CCP, cell cycle progression; CI, confidence interval; IQR, interquartile range; PSA, prostate specific antigen; RP, radical prostatectomy.

Table 3 Overview of the Cuzick et al. (2011) study

Study component

Description

Objectives/hypotheses

To assess the prognostic value of CCP in patients with prostate cancer.

Study design

Retrospective cohort.

Setting

1985–1995 Scott and White Clinic, US (RP cohort).

1990–1996 6 cancer registries in the UK (TURP cohort).

Inclusion/exclusion criteria

RP inclusion:

  • Patients who had RP for prostate cancer.

RP exclusion:

  • Patients who had been treated with neoadjuvant drugs.

  • Patients without clinical data and available tumour tissue.

TURP inclusion:

  • Men with clinically localised prostate cancer treated with watchful waiting.

  • Diagnosed following transurethral resection of prostate.

  • Under 76 years old at the time of diagnosis.

  • Had a baseline PSA measurement recorded.

TURP exclusion:

  • Patients treated with RP or radiation therapy within 6 months of diagnosis.

  • Patients who died or showed evidence of metastatic disease within 6 months of diagnosis.

  • Patients who had hormone therapy before the diagnostic biopsy.

Primary outcomes

Time to BCR for RP cohort; time to death for TURP cohort.

Statistical methods

Survival analysis was done with Cox proportional hazards models. The main assessment was a univariate analysis of the association between outcome and CCP score. A further predefined assessment of the added prognostic information after adjustment for the baseline variables was also done and a multivariate model was used.

Patients included

RP cohort:

n=410; median follow-up time 9.4 years (IQR 6.8–10.9); median age 68 years (IQR 63–71).

TURP cohort:

n=337; median follow-up time 9.8 years (IQR 5.4–11.8); median age 70.3 years (IQR 66.7-73.1).

Results

RP Cohort:

148/410 (36%) had BCR by 10 years after surgery. 366 scores were judged valid for statistical analysis.

The increase in hazard ratio for a 1-unit change in CCP score was 1.89 (95% CI 1.54 to 2.31; p=5.6×10−9). The multivariate analysis hazard ratio was 1.77 (1.40–2·22; p=4.3×10−6).

TURP cohort:

171/337 (51%) died within 10 years of diagnosis; 68 (20%) from prostate cancer and 103 (31%) from other causes.

The CCP score was the most important variable for prediction of time to death from prostate cancer in both univariate analysis (2.92, CI 95% 2.38 to 3.57, p=6.1×10−22) and the final multivariate analysis (2.57, 95% CI 1.93–3.43; p=8.2×10−11), and was stronger than all other prognostic factors, although PSA concentration also added useful information.

Conclusions

The CCP score was a good predictor of death from prostate cancer.

Abbreviations: BCR, biochemical recurrence; CCP, cell cycle progression; CI, confidence interval; IQR, interquartile range; PSA, prostate specific antigen; RP, radical prostatectomy; TURP, transurethral resection of the prostate.

Table 4 Overview of the Cuzick et al. (2012) study

Study component

Description

Objectives/hypotheses

To evaluate the clinical utility of the CCP score when generated from needle biopsies from men managed by watchful waiting.

Study design

Retrospective cohort.

Setting

6 UK cancer registries; 1990–1996.

Inclusion/exclusion criteria

Inclusion:

  • Men with clinically localised prostate cancer treated by watchful waiting.

  • Diagnosed using needle biopsy specimens.

  • Under 76 years old at the time of diagnosis.

  • Had a baseline PSA measurement recorded.

Exclusion:

  • Patients treated with radical prostatectomy or radiation therapy within the first 6 months after diagnosis.

  • Patients who died or showed evidence of metastatic disease within 6 months of diagnosis.

  • Patients who had hormone therapy before the diagnostic biopsy.

Primary outcomes

Death from prostate cancer.

Statistical methods

Survival analysis was carried out using a Cox proportional hazards model (time to death from prostate cancer). All p-values were 2-sided and 95% CI and p-values were based on chi-squared statistics with 1 degree of freedom, unless otherwise indicated, obtained from partial likelihoods of proportional hazards models.

A univariate analysis of the association between death from prostate cancer and CCP score was also performed.

Patients included

n=349 patients complete baseline and follow-up information; median age 70.5 years, median PSA 21.4 ng/ml, 91% Gleason score >7.

Results

Median CCP score was 1.03 with an interquartile range from 0.41 to 1.74. A 1-unit increase in CCP score was associated with a 2.02-fold increase in the hazard of dying from prostate cancer (χ2=37.6, p=8.6×10−10, 95% CI 1.62 to 2.53).

The 10-year death rate from prostate cancer was:

  • 19.3% for CCP score <0;

  • 19.8% for CCP score 0-1;

  • 21.1% for CCP score 1-2;

  • 48.2% for CCP score 2-3;

  • 74.9% for CCP score >3.

The multivariate analysis showed that extent of disease, age, clinical stage and use of hormone therapy were not statistically significant and therefore only CCP score, Gleason score and PSA level remained in the analysis. Multivariate analysis hazard ratio for CCP score was 1.65 (95% CI 1.31 to 2.09, p=2.6×10−5).

Conclusions

80% of the needle biopsies provided enough material to generate a CCP score. For these patients, the CCP score was a stronger prognostic factor than either the Gleason score or PSA levels.

Abbreviations: CCP, cell cycle progression; CI, confidence interval; PSA, prostate specific antigen.

Table 5 Overview of the Cuzick et al. (2015) study

Study component

Description

Objectives/hypotheses

To validate the prognostic value of a CCP score independently and in a pre-specified linear combination with standard clinical variables (the clinical CCR score).

Study design

Retrospective cohort study.

Setting

3 UK cancer registries; 2000–2003.

Inclusion/exclusion criteria

Inclusion:

  • Men aged under 76 years at diagnosis.

  • Men with clinically localised prostate cancer diagnosed by needle biopsy.

Exclusion:

  • Men treated with radical prostatectomy or radiation therapy within 6 months of diagnosis.

  • Men with objective evidence of metastatic disease (for example by bone scan, X-ray, radiograph, CT scan or MRI).

  • Men with clinical indications of metastatic disease (including pathological fracture, soft tissue metastases or spinal compression).

  • Men with a PSA measurement >100 ng/ml at or within 6 months of diagnosis.

  • Men who had hormone therapy prior to the diagnostic biopsy.

  • Men who died within 6 months of diagnosis or had <6 months of follow-up.

Primary outcomes

The prognostic value of the CCP score.

Statistical methods

Survival was analysed with a Cox proportional hazards model. The primary end point was time to death from prostate cancer.

A predefined combined CCR score encompassing both the CAPRA (linear) and CCP score was calculated to predict death from prostate cancer.

Further exploratory analyses included testing for proportional hazards, and testing for interactions of the CCP score with individual clinical covariates.

Patients included

n=761 (median age 70.8 years, IQR 66.5-73.6).

Results

In a univariate analysis, the CCP score hazard ratio was 2.08 (95% CI 1.76 to 2.46, p<6.0x10−14) for 1 unit change of the score.

In multivariate analysis including CAPRA, the CCP score hazard ratio was 1.76 (95% CI 1.44 to 2.14), p<4.2x10−7). The CAPRA score hazard ratio was 1.29 (95% CI 1.18 to 1.42; p<4.6x10-9).

The predefined CCR score was significantly predictive of death from prostate cancer, hazard ratio 2.17 (95% CI (1.83 to 2.57), X2=88.9, p<4.1x10−21).

Conclusions

The CCP score provides significant pre-treatment prognostic information and can be useful for determining which patients can be safely managed conservatively, avoiding radical treatment. The combined CCR score as a linear combination of the CCP score almost completely accounted for all molecular and clinical prognostic information.

Abbreviations: CAPRA, Cancer of the Prostate Risk Assessment; CCP, Cell cycle progression; CCR, Cell cycle risk; CI, confidence interval; CT, computerised tomography; IQR, interquartile range; MRI, magnetic resonance imaging.

Table 6 Overview of the Freedland et al. (2013) study

Study component

Description

Objectives/hypotheses

To evaluate the prognostic utility of the CCP score in men with prostate cancer treated with EBRT.

Study design

Retrospective cohort.

Setting

USA; 1991–2006.

Inclusion/exclusion criteria

Inclusion:

  • Patients who underwent diagnostic biopsy for prostate cancer and were treated with definitive EBRT.

Exclusion:

  • Patients without available formalin-fixed and paraffin-embedded blocks containing original diagnostic biopsy.

  • PSA level >100 ng/ml.

  • Patients who began treatment >2 years after diagnostic biopsy.

  • Patients with follow-up data for <3 years who had not developed BCR within the time frame.

Primary outcomes

Time to BCR event.

Statistical methods

Survival analysis was carried out using Cox proportional hazards models to assess the association between the CCP score as a continuous variable and risk of BCR. Most of the analyses are based on 5-year censoring to address the observed time dependence of HR for CCP.

Patients included

n=141; median age 66 years, IQR 60–71; 60% clinical stage T1; 61% Gleason score ≥7.

Results

The median CCP score was 0.12 (IQR –0.43, 0.66).

The HR for BCR was 2.55 (95% CI 1.43 to 4.55) for 1-unit increase in CCP score (p=0.0017).

The multivariable analysis included Gleason score, PSA, percent positive biopsy cores and androgen deprivation therapy; the HR per CCP unit was 2.11 (95% CI 1.05 to 4.25, p=0.034).

Conclusions

CCP was a significant predictor of BCR in patients having EBRT.

Abbreviations: BCR, biochemical recurrence; CCP, cell cycle progression; CI, confidence interval; EBRT, external beam radiation therapy; HR, hazard ratio; IQR, interquartile range; PSA, prostate specific antigen.

Table 7 Overview of the Crawford et al. (2014) study

Study component

Description

Objectives/hypotheses

To evaluate the impact of the CCP report on clinician treatment recommendations for patients with prostate cancer.

Study design

Prospective cohort.

Setting

USA; July to September 2013.

Inclusion/exclusion criteria

Inclusion:

  • Prostate cancer patients diagnosed by biopsy.

  • Patients who had CCP tests ordered by their clinician who completed both pre- and post-test report forms with intended selection of treatment.

Exclusion:

  • Not stated.

Primary outcomes

Binary change in treatment (a change from interventional to non-interventional therapy options) and the overall direction of change (to a more or less aggressive treatment).

Statistical methods

Outcomes were calculated along with their 2-sided 95% CI. The sample size was calculated to demonstrate a change of at least 10% (lower limit of 95% CI) in the magnitude of change between pre- and post-test recommendations assuming an observation of a 15% change in the study.

Patients included

n=331 patients, 67.4±7.43 years old. 82.5% had clinical stage T1c adenocarcinoma; 91.9% had Gleason scores of 6 or 7.

Results

The average CCP score was –0.69 with an average risk of 10-year mortality with conservative management of 3.5%.

Samples from 305 people were evaluable (in 26 people, the therapeutic decision was recorded as 'undecided' either pre-test or post-test). Overall, 64.9% (95% CI: 59.4 to 70.1%) showed a change between intended therapy options pre- and post-CCP test report.

There was a reduction in therapeutic burden in 40% of people (122/305), no change in 35.1% of people (107/305), and an increase in 24.9% of people (76/305).a

Conclusions

The use of CCP testing is associated with clinical utility among clinicians based on their changes in treatment plans for patients.

Abbreviations: CCP, cell cycle progression; CI, confidence interval.

a The therapeutic burden was defined by the following hierarchy: radical prostatectomy>radiation therapy>other therapy (brachytherapy/cryotherapy etc.)>androgen deprivation therapy>active surveillance>watchful waiting, where reduction in therapeutic burden includes both a shift from an interventional to a non-interventional therapy (from example from radical prostatectomy to active surveillance) as well as reduction in intended interventional burden (from example from radiation and radical prostatectomy to radiation only).

Table 8 Overview of the Shore et al. (2016) study

Study component

Description

Objectives/hypotheses

To evaluate the impact of the CCP test on shared treatment decision making for patients newly diagnosed with prostate cancer.

Study design

Prospective registry study with questionnaires.

Setting

USA; dates not specified.

Inclusion/exclusion criteria

Inclusion:

  • Men with recently (<6 months) diagnosed prostate cancer.

  • Men with histologically proven, presumed clinically localised prostate cancer.

  • Men who had not received any treatment and had sufficient biopsy tissue.

Exclusion:

  • Men with a known history of hypogonadism.

  • Men who had been treated with hormonal therapy.

Primary outcomes

Change in treatment.

Statistical methods

A subgroup analysis was conducted to assess change from interventional to non-interventional therapy options.

Multiple logistic regression was used to determine the impact of mortality risk, as determined by the CCP test, on treatment change.

Patients included

Of the 1,596 patients enrolled in the registry 1206 were eligible for analysis.

Mean age 65.9±8.36 years.

Results

There was a significant reduction in the treatment burden recorded at each successive evaluation (p <0.0001), with the mean number of treatments per patient decreasing from 1.72 before the CCP test to 1.16 in actual follow up.

The CCP test caused a change in actual treatment in 47.8% of patients. Of these changes 72.1% were reductions and 26.9% were increases in treatment burden. For every 1 unit increase in mortality risk there was an associated 2.7% increase in the odds of treatment increasing (and vice versa for decrease in treatment).

For each clinical risk category there was a significant change in treatment modality (intervention vs non-intervention) before compared with after CCP testing (p=0.0002).

Conclusions

The CCP test has a significant impact on shared decision making between patients and clinicians in terms of changes in treatment plans.

Abbreviations: CCP, cell cycle progression.

Table 9 Overview of the Warf et al. (2015) study

Study component

Description

Objectives/hypotheses

To demonstrate that the CCP score is a robust and reproducible molecular diagnostic tool that is appropriate for clinical use for the testing of either RP or needle biopsy FFPE samples.

Study design

The precision of the CCP score was assessed in a set of 6 biopsy and 12 RP samples.

Setting

All studies were performed within a CLIA-certified laboratory under established protocols.

Inclusion/exclusion criteria

The RP samples had sufficient tissue for 3 replicates, while the biopsy samples had sufficient tissue for 4 or 6 replicates. Samples were required to have mean expression of housekeeper (reference) genes ≤24 Ct, in order to match the average expression of clinical samples.

Primary outcomes

The analytical performance of the CCP test through assessment of:

  • Precision of the CCP gene expression signature.

  • Stability of stored RNA.

  • Yields of RNA extracted from FFPE tissue.

  • Linearity of the CCP score in relation to RNA concentration.

  • Amplification efficiency of genes within the CCP gene expression signature.

  • Dynamic range of the CCP gene expression signature.

Statistical methods

The precision for the overall CCP score was defined as the standard deviation captured in the residual variation term using a linear mixed model.

Samples included

6 biopsy and 12 RP samples.

Results

  • The overall SD of the signature was determined to be 0.1 CCP score units (95% CI, 0.08 to 0.13) between replicate measurements.

  • CCP scores were reproducible across all time points, with no trend in the scores of any of the individual samples

  • 100% of the RP and 99.8% of the biopsy samples produced sufficient RNA for testing

  • All samples had consistent CCP scores across the entire range of RNA concentrations that was assessed

  • None of the samples produced a CCP score at 0.06 ng/microlitre (1.5 ng of input RNA) because the CCP scores at those concentrations did not pass the quality control measures.

  • The linear range of the RNA concentration was from 62.5 to 0.24 ng/microlitre. This approximately 260-fold range exceeds the 20-fold range of RNA concentrations over which the signature was clinically validated and clinical samples are tested (40 to 2 ng/microlitre).

  • No statistical difference in the amplification efficiencies was observed when comparing housekeeper and target genes (p-value 0.39).

  • The observed range of the CCP scores was within recent clinical validations in prostate cancer samples (CCP scores from −2.0 to 4.1) and is well within the dynamic range of the gene expression signature.

Conclusions

The linear and dynamic range of the CCP signature exceeds the parameters utilized in clinical testing, indicating that the test is suitable for use.

Abbreviations: CCP, cell cycle progression; CI, confidence interval; Ct, cycle threshold; FFPE , formalin-fixed, paraffin-embedded; ng, nanograms, RNA, ribonucleic acid RP, radical prostatectomy; SD, standard deviation.

Table 10 Summary of the economic abstracts

Study

Country

Intervention (compared with standard treatment)

Population

Costs included

Original costs

Adjusted costs (PPP ER, inflation)

Crawford et al. (2015)

US

Prolaris

Men with localised prostate cancer (with 10 year follow up)

Costs of each unit of care that a patient might undergo (diagnostic, surgical, radiotherapy procedures and drug therapy)

$2,850 per patient, per year

£1,938

de Pouvourville (2015)

France

Prolaris

Men with localised low risk prostate cancer

Direct medical costs (for example drugs, staff time, and equipment)

At a hypothetical cost of €2,000 for the test, the lower limit of lifetime costs (discounted) is €1709 with an incremental gain of 0.23 QALYs.

An assumption of £1,502 for the test resulted in a discounted lifetime cost of £1,284 

Abbreviations: ER, exchange rate; PPP, purchasing power parity; QALY, quality-adjusted life year.