Introduction
Each year, approximately 17,000 Australian men diagnosed with prostate cancer (PCA) face a difficult decision whether to undergo treatment.1 This decision is difficult due to uncertainty about the risk of progression of cancer (and therefore the need for treatment), as well as the limited proven efficacy of PCA treatments in improving survival.2–4 Further complicating matters include the moderate risk of treatment-related side-effects impacting bladder, bowel and sexual function, potentially leading to adverse effects on physical, psychological and sexual well-being.5 Consequently, patients newly diagnosed with PCA are compelled to consider treatment options at a time of high psychological stress.6 Decision-making may therefore be significantly influenced by the patient’s preferences, which may be based on anxiety, fear, and the personal stories of family and other men, as well as more factual analysis, and possible cognitive decline.7
Patients diagnosed with PCA can undergo treatment with curative intent (TCI) with either surgery or radiation therapy, or treatment with non-curative intent (TNCI) including active surveillance, watchful waiting, or androgen deprivation therapy.8 The aim of managing PCA is to strike a balance, avoiding both under-treatment and over-treatment.9 Exact definitions of these terms are difficult and controversial,10 but Loeb et al suggest that over-treatment occurs when TCI is applied to a cancer detected through screening and that would not have been detected clinically or symptomatically in the patient’s lifetime.9
The imperative, therefore, is to identify and treat men with aggressive PCA who have a long life expectancy, thus avoiding under-treatment in these men.3 Moreover, clinicians must avoid over-treatment in men with limited life expectancy due to age, those with comorbidities who are likely to die from other causes, and those who have low-grade cancer and are therefore at low risk of PCA progression.11,12 Generally, younger men, with longer life expectancy and aggressive cancer have been shown to benefit significantly from TCI. However, there are also older otherwise healthy men with more aggressive cancers who may benefit from TCI but do not receive it.13–16
This study aimed to investigate the influence of cancer-related factors and patient-related factors (age and comorbidities) on the initial treatment decision for men with a new diagnosis of PCA. Treatment decisions were subsequently evaluated for alignment with the relevant guidelines and possible under- or over-treatment identified. This study is important as treating physicians need to carefully consider both cancer and patient factors when making treatment decisions if they are to avoid under- or over-treatment.
Methods and methods
The clinical records of 545 men who underwent prostate biopsy at four Australian-based urology services between January 2015 and December 2016 and received a first diagnosis of PCA were audited for this study. Two practices are located in metropolitan areas (n = 109 patients, two urologists), one in rural NSW (n = 99 patients, two urologists) and one in regional NSW (n = 337 patients, five urologists). The study was approved by the local Human Research Ethics Committee (HREC approval 2016/955).
Treatment
Patients were classified according to the treatment they underwent following their diagnosis of PCA. Patients allocated to the TCI group received surgery or radiation therapy. In this study, surgery refers to radical prostatectomy, either open or robotic-assisted.17 Radiation therapy refers to treatment with interstitial seeds (brachytherapy) or external beam radiation.
Patients in the TNCI group were allocated to receive active surveillance, watchful waiting, or androgen deprivation therapy (ADT).
Patient related factors
As part of the audit, patient demographic factors such as age, Charlson Comorbidity Index (CCI) and number of medications were recorded.18,19 Patient age was used to determine overall life expectancy based on the Australian Bureau of Statistics (ABS) life tables.20 The CCI was chosen because it is a validated tool for assessing comorbidities and is a strong predictor of overall survival and life expectancy.10,18 Additionally, the patients’ medical records were examined, and the number of regular medications prescribed was recorded as a numerical value.
Cancer related factors
Pathological details including serum prostate specific antigen (PSA) levels, cancer stage and tumor grade were collected from the prostate biopsy reports in the patients’ medical records. The clinical stage (using the 1992 American Joint Commission on Cancer (AJCC) staging system) was obtained either from the Digital Rectal Examination (DRE) findings recorded in the patient’s file or from the DRE findings recorded at the time of biopsy.21 The D’Amico risk group incorporates PSA, grade and clinical stage and guides to disease severity and prognosis.22 These risk groups correlate with biochemical recurrence, cancer specific survival and overall survival.23,24 Patients were classified into four risk groups according to the American Urological Association (AUA) and the American Society for Radiation Oncology (ASTRO) based on these D’Amico risk groups (Table 1a).23,25,26 Treatments received by patients in this study were compared with these guidelines to determine concordance.
Table 1Classification of patients into AUA/ASTRO Risk groups and the Lunardi et al algorithm to determine under- or over-treatment23
| D’Amico risk group | PCA grade group | PSA (ng/mL) | Clinical stage | Recommended treatment |
|---|
| (a) AUA/ASTRO risk groups |
| Low-Risk | Low | 1 | <10 | T1-T2a | AS |
| Intermediate-Favourable | Intermediate | 1 | 10 to <20 | ≤T2b-c | AS or TCI |
| Intermediate | 2 | <10 | ≤T2b-c | AS or TCI |
| Intermediate-Unfavourable | Intermediate | 2 | 10 to <20 | ≤T2b-c | TCI or WW |
| Intermediate | 3 | <20 | ≤T2b-c | TCI or WW |
| High-Risk | High | 4 or 5 | ≥20 | ≥T3 | TCI or WW |
| (b) Determination of under- or over-treatment (Lunardi et al algorithm)26 |
| Treatment | D’Amico risk group | Age | CCI | Over-treatment | Under-treatment |
| TCI | any | <75 | ≥2 | ✓ | |
| TCI | any | >75 | ≥1 | ✓ | |
| TCI | Low | 65–75 | ≤1 | ✓ | |
| None | Intermediate/ high | <75 | ≤1 | | ✓ |
| None | Intermediate/ high | 75–85 | 0 | | ✓ |
Lunardi et al.26 have developed an algorithm which considers both patient related factors (age and CCI) and cancer-related factors (D’Amico risk group) to determine under- or over-treatment of men with PCA (Table 1b).
Statistical analysis
Descriptive statistics were used to provide an overview of the respondents’ sociodemographic characteristics. To determine if the differences between the two groups were significant, when one value (the independent variable) was categorical, and the other was numerical or non-parametric (not normally distributed), the Mann- Whitney test was used as there were two categorical groups. When one value was numerical (and normally distributed) and the other was categorical, an unpaired t test was performed if there were two categorical groups. All analyses were performed using Prism 7 for MacOSX (GraphPad Software Inc.). Significance was set at a level of p < 0.05.
Results
The average age of the 545 patients included in this study was 67.5 years, with a range of 44 to 91 years (Table 2). The median PSA concentration was 8.0 ng/mL, ranging from 5.6 ng/mL in the Low-Risk group to 12 ng/mL in the High-Risk group. The average CCI for the entire cohort was 0.63, and the average number of medications was 2.9. Over 63.7% of the patients underwent TCI, and two-thirds of those underwent surgery. The percentage of patients receiving TCI did not vary based on geographical location (p = 0.709, Chi Square with Fisher’s exact test). Surgery was equally available in each of the practice settings. Specifically, surgery was chosen as the treatment option by 24% of patients from the metropolitan settings, 45% of patients in the regional settings and 34% of patients in the rural settings.
Table 2Audit of the factors affecting patient treatment choices following a diagnosis of PCA
| Total | Low risk | Intermediate favorable | Intermediate unfavorable | High risk |
|---|
| n (%) | 545 | 111 (20.4%) | 148 (27.2%) | 120 (22%) | 166 (30.4%) |
| Age (years, mean) | 67.5 | 63.1 | 66.7 | 68.9 | 70.3 |
| Average CCI | 0.63 | 0.55 | 0.51 | 0.57 | 0.85 |
| Number of medications | 2.9 | 2.5 | 2.8 | 2.7 | 3.3 |
| Median PSA | 8 | 5.6 | 6.9 | 9.5 | 12 |
| Treatment with Curative Intent (TCI) |
| Total | 347 (63.7%) | 29 (26.1%) | 114 (77%) | 89 (74.2%) | 116 (69.9%) |
| Surgery n (%) | 213 (39%) | 22 (19.8%) | 81 (54.7%) | 59 (49.2%) | 51 (30.7%) |
| Radiation n (%) | 134 (24.6%) | 7 (6.3%) | 32 (21.6%) | 30 (25%) | 65 (39.2%) |
| Treatment with non-curative intent (TNCI) |
| Total | 198 (36.3%) | 82 (73.9%) | 35 (23%) | 31 (25.8%) | 50 (30.1%) |
| Watchful waiting n (%) | 63 (11.6%) | 8 (7.2%) | 22 (14.9%) | 22 (18.3%) | 11 (6.6%) |
| Active surveillance n (%) | 88 (16.1%) | 74 (66.7%) | 10 (6.8%) | 4 (3.3%) | 0 (0%) |
| ADT n (%) | 47 (8.6%) | 0 (0%) | 3 (2%) | 5 (4.2%) | 39 (23.5%) |
| Concordance with recommended treatments |
| Concordance n (%) | 440 (80%) | 74 (66.7%) | 123 (83.1%) | 89 (74.2%) | 154 (94%) |
| Non-concordance n (%) | 104 (19%) | 37 (33.3%) | 25 (16.9%) | 31 (25.8%) | 11 (6.6%) |
| Non-Concordance, treated with Surgery | 22 (19.8%) | 22 (19.8%) | 0 | 0 | 0 |
| Non-Concordance, No treatment | 67 (64%) | 8 (7.2%) | 22 (14.9%) | 26 (21.7%) | 11 (6.6%) |
| Non-Concordance, Life expectancy <10 yrs | 15 | 0 | 4 | 8 | 3 |
| Lunardi estimate of over or under-treatment |
| Over-treatment n (%) | 77 (14%) | 10 (9%) | 22 (14.9%) | 15 (12.5%) | 30 (18.1%) |
| Under-treatment n (%) | 45 (8.3%) | 0 (0%) | 27 (18%) | 14 (12%) | 4 (2.4%) |
The distribution of patients across the four risk categories is detailed in Table 2. The AUA/ASTRO risk stratification was shown to be a primary determinant of treatment options for individual patients, with Low-Risk patients being more likely to have active surveillance and patients classified as Intermediate (Favorable and Unfavorable) or High-Risk being more likely to have TCI (surgery or radiation) (Fig. 1).
Treatment decisions for patients across all risk groups
Younger patients were more likely to be offered surgery (mean age 64.3 years) or active surveillance (mean age 61.8 years), and older patients were more likely to be treated with radiation, watchful waiting, or ADT. Patients treated with surgery were younger (mean age: 64.3 years) than those treated with all other treatment options combined (mean age: 69.7 years, p < 0.0001). Patients treated with surgery also had a significantly lower CCI (mean: 0.64) than all the other patients combined (mean CCI: 0.96, p < 0.0026) and were on fewer medications (mean number of medications 2.3) than were all the other patients combined (mean number of medications 3.6, p < 0.0001). However, overall, there was no significant difference in the age of patients receiving TCI compared to those receiving TNCI (Fig. 2a). Patients with more comorbidities (as indicated by a higher CCI) were more likely to receive TNCI (Fig. 2b, p = 0.0018). Similarly, patients on more medications were more likely to receive TNCI (P = 0.0442, Fig. 2c).
Treatment decisions according to the AUA/ ASTRO risk classification
As shown in Figure 1, patients classified as Low-Risk were most likely to undergo active surveillance (66.7% of this group, Table 2), which is in concordance with the AUA/ASTRO guidelines (Table 1a). Most patients who received a treatment not aligned with the AUA/ASTRO guidelines opted for surgery as a TCI option. Taking the Lunardi et al algorithm into consideration, 10% of our study population in the Low-Risk group received over-treatment (Table 2).26 However, all the patients, who received TCI (surgery or radiation) had a life expectancy of >10 years. Among patients classified as Low-Risk, there was no significant difference in age between those who received TCI (mean age 62 years) and those who received TNCI (mean age 63.5 years). Patients who received surgery had significantly lower CCI scores (mean 0.27) than those receiving active surveillance (mean 1.0, p < 0.0001).
Overall, 77% of the patients in the Intermediate-Favourable risk group were given TCI (55% surgery, 22% radiation therapy) (Table 2). Patients who received surgery were younger than those who received radiation therapy (mean age: 64 years for surgery versus 71 years for radiation, p < 0.0001), and those who received TCI were younger (mean age 66 years) than those who received TNCI (mean age 69 years, p = 0.0285, Fig. 3b). Patients who received TNCI were taking significantly more medications (mean 3.7) than those who received TCI (mean 2.5, p = 0.0288). Patients treated with surgery also had a lower average CCI (0.33) than those who received radiation therapy (0.87, p = 0.0350). However, overall, there was no significant difference in the mean CCI between those receiving TCI and TNCI. The AUA/ASTRO guidelines suggest the use of either the TNCI or the TCI for patients in the Intermediate-Favorable risk group (Table 1a), and there was high concordance in this group (83%, Table 2). According to the Lunardi et al algorithm, 15% of patients in this Intermediate-Favorable risk group were judged to have been overtreated either due to a high CCI or being over 75 years of age and still receiving TCI.26 In this group, 18% of patients (n = 27) were potentially undertreated.
Three-quarters of the patients in the Intermediate-Unfavorable risk group underwent TCI, with almost 50% receiving surgery and an additional 25% receiving radiation therapy (Fig. 1c). Similar to the previous risk group, patients who underwent surgery were younger (mean age 66 years) than those treated with radiation therapy (mean age 72 years, p = 0.0027). Those treated with TCI were younger (mean age 67.5 years) than those in the TNCI group (mean age 72.8 years), p = 0.0007, Fig. 3c). Those receiving TCI were taking fewer medications (mean 2.4) than those receiving TNCI (mean 3.6, p = 0.0128). Likewise, patients who received TCI had a lower CCI (mean 0.5) than those receiving TNCI (mean 0.9, p = 0.0145) and those undergoing surgery had a lower CCI (mean 0.3) than those receiving radiation (mean CCI 0.7, p = 0.0011) or watchful waiting (mean CCI 1.0, p < 0.0001). The AUA/ASTRO guidelines suggest the use of TCI for this group or TNCI if life expectancy is less than 5 years (Table 1a). Overall, three-quarters of patients received treatment in accordance with the guidelines (Table 2). According to the Lunardi et al algorithm, 12% of the patients (n = 14) in this risk group were undertreated, and 12.5% patients (n = 15) were considered overtreated because they received TCI despite having a high CCI score or being 75 years and older.26
Of the patients in the High-Risk group, 70% (n = 116) received TCI (31% surgery, 39% radiation) (Fig. 1d). Age influenced the treatment received, with patients receiving TCI being younger (average age 68.9 years) than those receiving TNCI (average age 73.7 years, Fig. 3d, p = 0.0032). Like patients in the intermediate risk groups, patients who underwent surgery were significantly younger (mean age 65 years) than those receiving radiation therapy (mean age 72 years, p < 0.0001). Patients who underwent surgery took fewer medications (mean 2.4) than those treated with radiation (mean 3.9, p = 0.0136).
The AUA/ASTRO guidelines suggest TCI or TNCI (watchful waiting) if the life expectancy is less than 5 years for patients classified as High-Risk (Table 1a). The treatment received by 94% of the patients in this High-Risk group was concordant with the guidelines (Table 2). According to the Lunardi et al algorithm, 18% of patients were potentially overtreated (Table 1b).
Discussion
The results of our study demonstrated that 80% of patients treated by the participating urologists received appropriate treatment based on the AUA/ASTRO guidelines. However, approximately one-third of patients classified as low-risk were not treated in accordance with the guidelines.25,27 Similar to the findings of the current study, a Victorian prostate cancer registry study, reported by Wang et al, showed a correlation between cancer risk stratification and treatment options.28 In their study, 55% of the low-risk patients received active surveillance, compared to 66% of the patients in our cohort. The most recent report of the Prostate Cancer Outcomes Registry of Australia and New Zealand which included more than 10,000 men diagnosed between 2015 and 2017, showed that, identical to our study, 66% of men with low-risk diseases were diagnosed via observation.29 In that study, 85% of men in the intermediate risk group received TCI compared to approximately 75% in our study. Our patient group had similar demographics to those in both of these studies. In particular, the risk stratification distribution was similar, suggesting that our patient cohort is representative of Australian patients with PCA in general.28,29 The results of these studies and ours confirm that patients’ treatment options are initially being appropriately influenced by cancer factors.
Once treatment options are established based on cancer factors, patient age, and comorbidities, these factors will modify the available options.8,30 Across all risk groups, in our study, younger and healthier patients were more likely to undergo TCI, indicating that invasive treatments were given to those most likely to benefit. In the current study, patients classified in the Low-Risk group receiving TCI had fewer comorbidities than those receiving TNCI. Similarly, for patients classified into the High-Risk group, those receiving surgery were on fewer medications than those receiving other therapies, including radiation (a second TCI). These findings concur with an Australian population-based study of men with a new diagnosis of PCA, which showed that men who were younger (<60 years old) and had fewer comorbidities were more likely to receive surgery (radical prostatectomy).31 Hence, it is critical to estimate the risk of death from PCA and compare it with the risk to the risk of death from other causes to avoid utilizing TCI in men with limited life expectancy or significant comorbidities.32 Schymura et al.’s study supports this paradigm of treatment with TNCI associated with increasing age, high PSA, low grade cancer and high comorbidity.33
Under-treatment should be avoided for men who are young, otherwise healthy, and have aggressive cancer by offering TCI. Similarly, over-treatment (unnecessary treatment) should be avoided in patients with low-grade cancer, or who are likely to die from competing causes, rather than from PCA because of limited life expectancy due to significant comorbidities. However, older, otherwise healthy men with aggressive cancer may need to receive treatment.29 In our study, over-treatment was identified in 15% of patients, which is slightly lower than the over-treatment rate of 25% reported by Lunardi et al.26 In the current study, like in the Lunardi study, patients identified as being over-treated received TCI despite having a limited life expectancy and/or high incidence of comorbidities. Daskivich et al.’s study of nearly 1,500 patients with newly diagnosed low/intermediate risk PCA showed that the risk of death from PCA over 10 years was between 5- and 8% (regardless of the treatment chosen), with 25% of men dying from non-PCA causes.10 Therefore, for many men, the risk of dying directly from PCA is low, while those with a CCI greater than or equal to 2 have a >75% risk of dying from any cause over 10 years.3,27 Ultimately, comorbidities have a significant impact on survival, according to Frendl’s study showing that age at diagnosis, CCI score, self-reported general health and smoking are the most predictive risk factors for mortality.14 Scores utilizing medication use have been shown to be a good measure of comorbidity.34,35
The impact of age on treatment choice following a diagnosis of PCA is complex. Research has established that older individuals, males, and those with lower education levels are more likely to prefer a more passive role in treatment decisions, potentially hindering their full participation in the decision-making process.36 This could affect many men with prostate cancer. Moreover, increasing age has been associated with the tendency to seek less information, make decisions more quickly, prefer fewer options, have increased difficulty understanding information, and place greater emphasis on emotional aspects when making decisions.37 In addition, some patients wish to avoid having to make choices and therefore rate physician advice as the most important factor in treatment decision-making.38,39 Importantly, some physicians maintain a paternalistic view of decision making and may not fully explore the patients’ wishes,40 despite the evidence suggesting that greater patient involvement in the decision-making process leads to increased patient satisfaction with the decision.41,42
Age is a significant independent risk factor for comorbidities and plays a pivotal role in determining treatment options.12 In our study, patients receiving TCI were significantly younger than those choosing TNCI for the Intermediate-Favorable, Intermediate-Unfavorable and High-Risk groups. Clinicians can use ABS life tables as a crude estimate of life expectancy based on age.43 However, these tables do not consider health status or comorbidities and have been shown to overestimate life expectancy significantly, especially in older men, which may contribute to possible overtreatment.11,12 Despite this evidence, Daskivich et al. suggested that clinicians tend to emphasize age (and crude estimates of life expectancy) over comorbidities when making treatment decisions.10 Hoffman et al. confirmed that comorbidity is a more significant determinant of life expectancy than age.11,16
However, until recently, the focus of the discussion has been on over-treatment of older men with co-morbidities. It is crucial to recognize that some older men with aggressive cancer may be unjustly denied TCI based merely on their age, and thus suffer from under-treatment. For older men with high grade disease, the lethality of PCA should not be underestimated, especially for those with fewer comorbidities.16 In our study, 12–18% of men in the intermediate risk groups did not receive TCI despite having no comorbidities suggesting possible undertreatment of this group of men. Lu-Yao showed that men with high-risk diseases and a life expectancy greater than 10 years face a significant risk (>25%) of PCA specific mortality within that 10-year period, suggesting the potential benefits of active treatment.13 However, men over 70 years are less likely to receive TCI, regardless of cancer grade or CCI, suggesting that age alone, rather than cancer grade or comorbidities, is the prime determinant of treatment options for these individuals.36 In Frendl et al’s study, for men over 65 years, only 40% of those who died due to PCA had undergone definitive treatment, indicating potential undertreatment in this older age group.14
Conclusions
The aim in treating men diagnosed with PCA should be to avoid under-treatment in men who are young, healthy, and have aggressive cancer by offering TCI. Conversely, over-treatment or unnecessary treatment should be avoided in men with low-grade cancer or in those who are likely to die from competing causes rather than from PCA because of limited life expectancy or significant comorbidities. The results of the current study demonstrated that 80% of patients treated by the participating urologists received appropriate treatment according to the AUA/ASTRO guidelines. Ultimately cancer factors are the prime determinants of treatment options and can be modified by life expectancy and comorbidities. Physicians must be careful not to overestimate the lethality of cancer and underestimate the potential for age and comorbidities to be more likely causes of death while still offering TCI to those most likely to benefit.
Abbreviations
- ADT:
androgen deprivation therapy
- ASTRO:
the American Society for Radiation Oncology
- AUA:
American Urological Association
- CCI:
Charlson comorbidity index
- DRE:
Digital rectal examination
- PCA:
prostate cancer
- PSA:
prostate specific antigen
- TCI:
treatment with curative intent
- TNCI:
treatment with non-curative intent
Declarations
Acknowledgement
Julie Edwards RN. and Samantha Ognenovska for their work on data collection and entry. Also thanks to Dr Nicholas Thorpe for his work on the initial interpretation of the data.
Data sharing statement
The complied audit data used to support the findings of this study are included within the article; The data relating to individual patients used to support the findings of this study have not been made available because of restrictions associated with ethics approval.
Ethical statement
This study was approved by the Joint University of Wollongong/ Illawarra Shoalhaven Local Health District Human Research Ethics Committee (HREC approval 2016/955). The study conformed with the ethical guidelines of the Helsinki Declaration (as revised in 2013). As this study was an audit of clinical notes, approval from each individual patient was not required.
Funding
No funding was available to support this study.
Conflict of interest
The authors have no conflict of interests related to this publication.
Authors’ contributions
Study concept and design (TAS, JRM, KJM), acquisition of data (TAS, WW, SN), analysis and interpretation of data (TAS, WW, SN, KJM), drafting of the manuscript (TAS, JRM, KJM), critical revision of the manuscript for important intellectual content (JRM, KJM) and study supervision (KJM). All authors have made a significant contribution to this study and have approved the final manuscript.