Background
Prostate cancer is a major public health issue. In developed countries, it is one of the most common male cancers. Reference Makino, Sato and Takenaka1,Reference Giordano, Lee and Kuo2 In Africa, and more specifically in Gabon, it tops the list of the most common male cancers, with 20·5% of new cases recorded in 2022. This incidence ranks it alongside liver cancer and colorectal cancer. It is estimated that one in ten men is at risk of developing this disease before the age of 75. The mortality rate is also high, reaching 6.8% for this same age group, which positions it as one of the leading causes of cancer-related deaths, on par with liver cancer and colorectal cancer. For all populations combined, it is the fourth leading cause of cancer in terms of new cases (8·1%) and deaths (6·3%). Reference Ferlay, Ervik and Lam3
External beam radiotherapy (EBRT) is a major therapeutic option against this disease, either alone or in combination with hormone therapy. Reference Viswanathan, Yorke and Marks4 Despite its benefits for the tumour, it can still cause acute and/or late genitourinary toxicities. Reference Rancati, Palorini and Cozzarini5 Acute symptoms typically appear during or shortly after treatment, within three months. The most common include pollakiuria and urinary urgency, dysuria, nocturia, haematuria, and interrupted or weak urine flow. Late symptoms can occur several months, or even years, after the end of treatment. These late symptoms include urethral narrowing (urethral stricture), which makes urination difficult, urinary incontinence and haemorrhage. Eddy Valgueblasse, Aurélien Descazeaud and the French Urology Association’s Committee on Male Micturition Disorders (CTMH-AFU) report that the risk of radiation cystitis is significant, potentially causing irritative signs in 50% of cases (pollakiuria, nocturia, urgency). Reference Valgueblasse and Descazeaud6 The study by Grise et al. Reference Grise and Thurman7 shows that the incidence of late urinary incontinence is less than 10%.
Radiation-induced urinary toxicities or symptoms can significantly impair patients’ quality of life. A proactive approach to these risks is essential to minimise the long-term impact.
With this in mind, Nikitas et al. Reference Nikitas, Jamshidian and Tree8 emphasise in their study that it is possible to minimise radiation-induced toxicities by reducing aggressive margins around target volumes using MRI guidance.
Identifying and monitoring these symptoms allows healthcare professionals to adapt treatment or offer early interventions. The challenge of radiotherapy for prostate cancer, therefore, is to achieve a differential effect between the tumour cells in the prostate and the surrounding healthy cells. While new irradiation techniques like intensity-modulated radiotherapy (IMRT), image-guided radiotherapy (IGRT), and volumetric modulated arc therapy (VMAT) improve local tumour control and reduce toxicity to organs at risk compared to 3D-CRT, Reference Catucci, Alitto and Masciocchi9 understanding the impact of dosimetric and clinical factors on urinary symptoms after radiotherapy is crucial for helping with decision-making for a more tailored treatment and management plan for patients. In an ideal world, all radiotherapy treatments in the near future would be personalised, taking into account not only the patients’ clinical characteristics but also different races and potential side effects. The development of decision support systems to personalise treatments is therefore a key issue in oncology. Reference Pisani, Galla and Loi10 The study by Ratnakumaran et al. Reference Ratnakumaran, Hinder and Brand11 highlights the importance of an individualised approach, from treatment planning to long-term follow-up. Taking initial symptoms and acute toxicity into account can help prevent or better manage late urinary complications, thereby improving patients’ quality of life.
It is against this backdrop that a study focused on predicting urinary toxicity following curative treatment for localised and locally advanced prostate cancer, specific to our patients, becomes relevant. The aim of this retrospective study was therefore to identify predictive factors for acute and late genitourinary toxicities based on clinical and dosimetric data from patients who received curative prostate cancer treatment between 2013 and 2021 at the Akanda Cancer Institute, Gabon.
Methods
This section outlines the methodology, techniques and procedures used to achieve our goals. It begins with patient selection, moves on to treatment characteristics, treatment planning and treatment delivery, then moves on to patient follow-up and concludes with statistical analyses that were used in this study.
Patient selection
We retrospectively analysed the files and treatment plans (from 2013 to 2021) of 72 patients with clinically staged cT1c-T4N0-1M0 prostate adenocarcinoma treated at the Department of Radiotherapy and Medical Physics at the Akanda Cancer Institute in Gabon. From the original cohort, 26 patients were excluded because their medical files were incomplete or they had not been followed up at the Akanda Cancer Institute. Finally, the study population thus consisted of 46 patients. Prior to any therapeutic approach, patients were informed of the treatment regimen and possible side effects, and verbal consent was obtained before treatment. The treatment regimen was a combination of hormone therapy (LH-RH analogue) and 3D conformal radiotherapy (3D-CRT) for all patients. The full patient characteristics are listed in Table 1 below.
Patient and disease characteristics
PSA = prostate-specific antigen (ng/mL); statistical parameters: age at diagnosis, comorbidity, TNM stage, D’Amico Risk, prescription dose (Gy). N: sample size; %: percentage.
Treatment characteristics
In the supine position, computed tomography (CT) images were acquired using a Philips Brilliance Big Bore CT scanner (16 cuts, 85 cm ring and 60 cm field of vision). Bladder delineation (whole organ) and shape depend on repletion; that is the reason why, before the CT scan, patients were asked to empty their bladder and drink a comfortable volume of water not exceeding 1.5 L (on average 800 mL of water). Tomography slices of 3 to 5 mm were acquired. The entire prostate target volume was defined. No lymph node dissection was performed on the single patient who underwent surgery. The estimation of lymph node invasion was established according to Roch’s formula 12 ; the operated patient (prostatectomy) benefited from irradiation of the lymph node areas according to the results of the calculation. A 10 mm margin was applied in all dimensions. For the prostate and vessels, a margin of 3 to 5 mm was used.
Treatment planning
The treatment plan was calculated using the CMS-Xio treatment planning system (TPS). The treatment procedure was divided into two or three sequences. Irradiation to Planning Target Volume (PTV), i.e., PTV1, PTV2 and PTV3. In the first sequence, a dose of 46 Gy in 23 fractions was delivered to PTV1 using the four beams from the left (270°), right (45°), anterior (0°) and posterior (180°). In the second sequence, a median dose of 11 Gy (range: 8–14 Gy) was delivered to PTV2 using six beams in 225°, 270°, 315°, 45°, 90° and 135° field directions. In the third sequence, a median dose of 18 Gy (range: 12–22 Gy) was delivered to the PTV3 using six beams in the same orientation as sequence 2. The median prescription was 74 Gy/37 fractions/7.5 weeks. The prescription point was set to the isocentre. The treatment plan was designed to encompass the PTV with 95% of the prescribed isodose line as long as the rectal dose was acceptable for the criteria. All patients were treated with full 3D-CRT with a median dose of 74 Gy (range: 66–80 Gy). Reference Mabika Ndjembidouma, James and Ondo Meye13 The dose constraints applied during bladder dosimetry planning were either V60 ≤ 50% or V65 ≤ 50%, V70 ≤ 35%, V75 ≤ 25% and V80 ≤ 15%. The dosimetric parameter Vx (V60, V65, V70, V75) represents the volume of the bladder receiving x Gy. It is the most objective dosimetric criterion for validating a treatment plan. Knowing these dose volumes makes it possible to assess the occurrence or not of radiation-induced toxicity.
Treatment
The treatment was administered with a 15 MV photon beam using an Elekta Precise Linear Accelerator (LINAC). The treatment plans were generated using a conformal radiation therapy technique on an XIO (Elekta CMS) TPS. These procedures are described in detail in our preliminary study. Reference Mabika Ndjembidouma, James and Ondo Meye13
Follow-up
Patients were evaluated weekly during the treatment for acute GU adverse events. After 3D-CRT, physical examinations with prostate-specific antigen (PSA) measurement were performed every 3 months during the first two years and every 6 months thereafter. Reference Mabika Ndjembidouma, James and Ondo Meye13 GU complications were graded according to the Common Terminology Criteria for Adverse Events (CTCAE v4.0).
Statistical analysis
For statistical analysis, variables were categorised as either qualitative (e.g., comorbid conditions such as diabetes, arterial hypertension, pre-existing urinary symptoms, history of catheterisation and tumour characteristics) or quantitative (including the age of patients and all dosimetric parameters). To analyse and to find a potential correlation between toxicity and qualitative variables, the Chi-squared and Fisher’s exact tests were used to examine whether there was a statistically significant association in the occurrence of acute and late GU toxicity. To find the relationship between dosimetric parameters with acute and late bladder adverse events, the Mann–Whitney test was applied. To determine the predictive factors of acute and late GU toxicities following irradiation for localised and locally advanced prostate cancer in our institution, we applied logistic regression. A p-value of <0.05 was considered statistically significant. Analyses were performed using SPSS 21 software. To evaluate the goodness of fit of the model, we used the Hosmer–Lemeshow test. The proportion of the variance of the dependent variable (acute and late GU toxicities) explained by the independent variables (dosimetric parameters and patient characteristics) was estimated using Nagelkerke’s R2. The model’s performance was evaluated using classification metrics derived from the confusion matrix, namely precision, recall and the F1 score, according to the following mathematical formulations:
Precision = TP/(FP + TP)
Recall = TP/(FN + TP)
F1 score = 2 × Precision × Recall/(Precision + Recall).
Results
After a median follow-up of 57.5 months (range: 39 - 88 months), no local recurrences or prostate cancer-related deaths were observed during the follow-up period. All patients remained disease-free during the follow-up. Overall survival related to prostate cancer was therefore 100%. No patients interrupted treatment due to acute side effects. Reference Mabika Ndjembidouma, James and Ondo Meye13
Toxicity incidence
In our cohort, only one patient underwent surgery and did not experience any post-radiotherapy toxicity, either acute or late. A summary of acute and late GU side effects is shown in Table 2.
Incidence of acute and late genitourinary (GU) adverse events
GU adverse events related to the Common Terminology Criteria for Adverse Events (CTCAE v4.0).
Statistical parameter: side effect. N: sample size; %: percentage.
In this study, 50% (23/46) of the patients experienced acute GU side effects. The majority of these were low-grade, with Grade 1 acute GU complications observed in 37·0% (17/46) and Grade 2 complications in 13·0% (6/46) of the cohort, respectively. Critically, no patient developed acute complications exceeding Grade 2. According to the CTCAE v4.0, toxicity endpoints were cystitis, dysuria, haematuria and pollakiuria. In total, 11 patients (23·9%, 11/46) developed Grade 1 cystitis, and 13% (6/46) of patients developed grade 2 cystitis. Grade 1 haematuria was observed in 4·3% (2/46) of patients. Grade 1 and 2 pollakiuria were experienced in 8.7% (4/46) and 4·3% (2/46) of cases, respectively. Grade 1 and 2 dysuria were seen in 13·0% (6/46) and 2·2% (1/46) of the patients, respectively. The majority of acute GU side effects disappeared a few weeks after radiotherapy. No patient developed late cystitis, late dysuria or late pollakiuria of grade 2.
The median time to the development of late GU Grade ≥1 toxicities was 22 months (range: 9–76 months). In our study, 80·4% (37/46) did not present any late GU toxicity. Only 19·6% (9/46) experienced late GU side effects. Among those, Grade 1 dysuria was observed in 6·5% (3/46), haematuria in 8·7% (4/46) and pollakiuria in 8·7% (4/46). There were no late grade 2 or higher complications.
Toxicity assessment
Assessment of GU toxicities and clinical parameters
Table 3 below shows the possible correlation between GU toxicity and several clinical factors, including patient age, comorbid conditions (diabetes and arterial hypertension), pre-existing urinary symptoms, history of catheterisation and tumour characteristics (Gleason score, Tumour Node Metastasis (TNM) stage, and Amico risk), as well as the prescribed dose.
Correlation between acute and late GU toxicity with patient characteristics
GS = Gleason Score. Statistical parameters: age at diagnosis, hypertension, diabetes, clinical stage, Gleason score, D’Amico Risk, prescription dose (Gy). N: sample size; %: percentage. Values are considered statistically significant when p < 0.05.
The assessment of genitourinary toxicities in relation to patient characteristics revealed a statistically significant association (p = 0·022) between hypertension and late urinary toxicity. The Gleason score was also correlated with late genitourinary toxicities. Compared to other Gleason scores, patients with a 7b (4+3) score were more likely to experience late side effects. The chi-square test indicated that this difference was statistically significant (p = 0·012). No statistically significant association with the other parameters tested was observed.
Assessment of pre-existing urinary symptoms and the occurrence of acute and late GU toxicities
Before radiotherapy, some patients had urinary symptoms such as pollakiuria, dysuria and haematuria. Others had pollakiuria associated with haematuria or dysuria and complicated pollakiuria due to acute urinary retention. For these pre-treatment functional signs, we also sought to establish a possible correlation between the presence of these symptoms, the history of urinary catheterisations and the occurrence of acute and late GU toxicities following radiotherapy of localised prostate cancer (Table 4).
Evaluation of pre-existing urinary symptoms and the occurrence of acute and late GU toxicities
Singles: ‘+’ = associated with; ‘→’ = following of.
The chi-square test revealed that the pre-treatment presence of a combination of symptoms (pollakiuria and haematuria, or pollakiuria and acute urinary retention) showed a statistically significant association with late GU. A similar correlation was found for patients with a history of urinary catheterisation. These correlations were statistically significant (p = 0·04 for the two symptom pairs and p = 0·044 for the history of urinary catheterisation). Cramer’s V, which quantifies the strength of the relationship between the correlated parameters, is 0·302 for the first two correlated symptoms. The probability of developing late GU toxicity for patients with pollakiuria associated with haematuria, or pollakiuria followed by Acute Urinary Retention (AUR), was 30·2% for our cohort. The Cramer’s V associated with a history of urinary catheterisation and late GU toxicity was 0·297, indicating a relationship strength of 29·7% in our study.
Relationship between acute and late side effects
Finally, we assessed whether the onset of acute GU toxicity was associated with the occurrence of late GU toxicity (Table 5).
Assessment between acute and late side effects
The chi-square test revealed a statistically significant association between acute and late GU toxicity (p-value = 0·009). Patients who developed acute toxicity were more likely to experience late urinary toxicity (38·4%), with an associated Cramer’s V of 0·384.
Acute and late GU toxicity assessments with irradiation parameters
We also investigated the relationship between the occurrence of acute and late GU toxicities and dosimetric parameters. The results are listed in Table 6.
Correlation between acute and late GU toxicity with dosimetric parameters
V = urinary bladder volume delineated; D mean = mean dose received by the urinary bladder; D max = maximum dose received by the urinary bladder; V x = percentage of urinary bladder volume receiving x Gy; D x,V = urinary bladder volume irradiated at x Gy; D T,V = bladder volume irradiated at the prescribed dose; SD = standard deviation; Prostate V = volume of prostate gland, PTVprostate = clinical prostate volume. Statistical parameter: dose-volume parameters; N: sample size; %: percentage. Values are considered statistically significant when p < 0.05.
The Mann–Whitney showed that the occurrence of acute urinary toxicity was correlated with the average dose (p = 0·042), the prostate gland volume (p = 0045), the planning prostate volume (p = 0·029) and the percentage of bladder volume irradiated at the doses of 60 Gy (p = 0·017), 65 Gy (p = 0·042) and 70 Gy (p = 0·012), respectively. No correlation between late urinary toxicity and irradiation parameters was observed.
Toxicity prediction
The preliminary analyses of the correlates showed a significant distinction between the risk factors for acute and late genitourinary toxicities. Acute GU toxicities showed a correlation with quantitative variables, particularly dosimetric parameters, whereas late GU toxicities were uniquely associated with patient-specific factors. Logistic regression identified the variables within these two major groups that are independent predictors of acute and late GU toxicity. Multivariable logistic regression for predicting acute GU toxicity is presented in Table 7 below.
Multivariable logistic regression to identify covariables associated with acute GU
DT = irradiated dose or prescribed dose.
The results of our multivariate logistic regression model show that prostate gland volume is predictive of acute genitourinary toxicity (OR 1·132, CI (1.018–1.260), p = 0·022); clinical prostate volume was also found to be a predictor (OR 0·944, CI (0·899–0.991), p = 0·021).
The Hosmer–Lemeshow test result (p = 0·2) indicates a good fit of the model with the data. The result of Nagelkerke’s R2 was 0·543, which implies that 54.3% of the variability in toxicities can be explained by the quantitative parameters of the model. We obtained an accuracy of 0.857, a recall of 0·818 and an F1 score of 0·837. These values demonstrate that our logistic regression model is robust and reliable in its predictions.
The multivariable logistic regression for predicting late GU toxicity is presented in Table 8.
Multivariable logistic regression to identify covariables associated with late GU
The results of the second logistic regression indicated that arterial hypertension (OR 7.755, CI (1·105–54·403), p = 0·039) and the presence or onset of GU toxicity during treatment (OR 13.243, CI (1.257–139·499), p = 0·032) are predictive of late UG toxicity. The non-significant Hosmer–Lemeshow chi-squared test (p = 0·858) indicates that our second logistic regression model fits well with the observed data. The Nagelkerke R2 was 0·455, which explains 45·5% of the observed variability in toxicities. The metric results demonstrate the overall robustness of our model, with an accuracy of 0·892 and an F1 score of 0·826. However, the low recall value of 0·556 is related to the fact that we removed some variables due to their high multicollinearity or complex interaction with other parameters.
Discussion
Urinary toxicity or genitourinary toxicity includes various symptoms such as cystitis, haematuria, urinary frequency and dysuria, and the clinical severity of these toxicities is classified using the Common Terminology Criteria for Adverse Events (CTCAE v4.0).
Analysis and comparison of genitourinary toxicity
In the literature, the incidence of acute GU toxicities is around 40% and late 5 to 13% in the context of radiotherapy for pelvic cancers. Reference Rigaud, Hetet and Boucho14,Reference Nakamura, Monti and Castilho15 These incidences are comparable to those observed in our study, which was 50% (23/46) for acute toxicities and 19·6% (9/46) for late toxicities. Table 9 below summarises the incidence of GU toxicities identified in the literature.
Incidence of acute and late GU toxicities in the literature
Although half of our patients developed acute GU toxicities, these were predominantly Grade 1 (37%, 17/46) and Grade 2 (13%, 6/46). This observation could explain the non-interruption of treatment for all patients. Regarding late GU toxicities, our study population did not develop toxicities greater than Grade 1, which contrasts with the incidence observed in the literature. This justifies the fact that the post-radiation quality of life of our patients was not altered.
Having highlighted the incidence of GU toxicities following localised treatment of prostate cancer in patients irradiated and followed at the Akanda Cancer Institute, we proceeded to perform statistical tests to establish possible correlations between the appearance of these GU toxicities and clinical and/or dosimetric parameters.
Clinical factors and genitourinary toxicity
The development of side effects after anticancer treatment is intrinsically multifactorial and cannot be explained by dosimetric factors alone. Reference Jackson, Marks and Bentzen24 De Crevoisier et al. Reference De Crevoisier, Fiorino and Dubray25 similarly highlighted that the risk of urinary toxicity appears more linked to patient-specific risk factors than to irradiation dose-volume parameters. They reported that the risk factors for late urinary toxicity identified in the literature included pre-existing irradiation, drug treatment for pre-existing urinary symptoms, a history of transurethral resection, age over 70 years, diabetes and hormonal treatment. For many authors, clinical parameters, habits and lifestyle are also predictive of acute and late urinary toxicity. Reference Peeters, Heemsbergen and van Putten19,Reference Wortel, Incrocci and Pos21
From our study and in accordance with our logistic regression model, arterial hypertension (OR 7·755, CI (1·105–54·403), p = 0·039) is predictive of late genitourinary toxicity. Other authors, however, have demonstrated that the use of antihypertensives is predictive of acute urinary toxicity, Reference Cozzarini, Rancati and Carillo26,Reference Palorini, Rancati and Cozzarini27 a finding which could be applicable to our series. In fact, antihypertensive drugs can increase the production of free radicals, which are unstable molecules that can damage cells. Radiotherapy can also increase oxidative stress, and the combination of the two can lead to significant cell damage in healthy tissues, including the bladder. Additionally, some antihypertensive drugs, such as calcium channel blockers and direct vasodilators, are powerful peripheral vasodilators that reduce blood pressure by lowering total peripheral vascular resistance. Reference Bakris28 This effect could consequently improve tissue oxygenation, which is a key factor for the effectiveness of ionising radiation on both healthy tissues and tumour cells (via tumour perfusion).
However, current data are insufficient to conclude on the major deleterious effects of antihypertensive drugs, but they justify an individualised approach that considers the expected radiation-induced toxicity profile for each patient.
The impact of pre-existing urinary symptoms and catheterisation on genitourinary toxicity
Richaud et al. Reference Richaud, Moreau and Eschwege29 underscore that the relationship between pre-existing urinary symptoms, a history of urinary catheterisation, and the occurrence of acute and late GU toxicities following radiotherapy for localised prostate cancer is not widely described in the literature. Nevertheless, it is a standard clinical guideline not to irradiate a very large adenomatous prostate in patients with a high International Prostate Symptom Score (IPSS); otherwise, a retention syndrome may occur during treatment. Urinary incontinence, which can manifest as urge or be associated with dysuria, a side effect secondary to irradiation of the prostatic urethra and bladder base, is rare, yet it may be aggravated by pre-radiotherapy bladder dysfunction or urinary obstructive syndrome. Reference Richaud, Moreau and Eschwege29
The preliminary univariate study in our series clearly demonstrated a relationship between specific pre-existing urinary symptoms (for example, pollakiuria associated with haematuria, and pollakiuria followed by acute urinary retention) and a history of urinary catheterisation with the onset of late genitourinary toxicity. These relationships were statistically significant for the presence of urinary symptoms (p = 0·040) and for a history of urinary catheterisation (p = 0·044). Importantly, our preliminary results align with those reported by Ratnakumaran et al. Reference Ratnakumaran, Hinder and Brand11 ; their univariate analysis, specifically, demonstrated that acute obstructive urinary symptoms of grade 2+ (p < 0·0001), irritative symptoms (p = 0·003), and incontinence (p = 0·01) were all separately associated with late urinary toxicity. Our multivariate logistic regression model, however, failed to find a significant predictive relationship between these variables and the onset of late genitourinary toxicities.
The analysis of self-reported symptoms conducted by Pisani et al. Reference Pisani, Galla and Loi10 showed a significant correlation between IPSS baseline values (p = 0·009), nocturia (p = 0·002), bladder urgency (p = 0·024) and incontinence (p = 0·024) and the development of acute urinary toxicity in their univariate analysis. Crucially, their subsequent multivariate logistic regression analysis demonstrated that IPSS value, nocturia and urinary incontinence retained a significant correlation with acute toxicity (p = 0·0003).
Acute genitourinary toxicity as a predictor of late urinary complications
In patients who developed GU toxicity during radiotherapy or within three months of irradiation, our study showed that acute GU toxicity was predictive of late urinary toxicity. This correlation is statistically significant (p = 0·009). This observation was also documented in the study by Taleb et al. Reference Taleb, Al and Boukerche30 in their series, demonstrating that the majority of patients who had developed a late urinary complication had already experienced acute toxicity during radiotherapy. Zelefsky et al. Reference Zelefsky, Levin and Hunt31 reported in their study that acute urinary symptoms that occurred during radiotherapy were associated with an increased risk of late grade 2 urinary symptoms. They showed that the existence of acute urinary symptoms multiplied the risk of late sequelae by 3·5 times. The studies by Ratnakumaran et al. Reference Ratnakumaran, Hinder and Brand11 and Nikitas et al. Reference Nikitas, Jamshidian and Tree8 confirm that acute toxicity is an important predictive variable for late GU toxicity after localised prostate radiotherapy using Stereotactic Body Radiation Therapy (SBRT) and conventional or hypofractionated schemes. This is in direct alignment with our own findings. In our series, the multivariable logistic regression analysis showed that a grade ≥1 acute GU toxicity was significantly associated with developing a grade ≥1 late GU toxicity after 3D-CRT (OR 13·243, 95% CI (1.257–139·499), p = 0·032).
The role of bladder dosimetry and prostate volume in acute and late genitourinary toxicity
Regarding irradiation parameters, Pisani et al. Reference Pisani, Galla and Loi10 documented in their study that bladder doses greater than 75 Gy could be predictive of acute urinary toxicity. Michalski et al., Reference Michalski, Purdy and Winter32 on the other hand, found a statistically significant correlation (p = 0·013) between the dose-volume parameter of the bladder if V65 > 30% with the increase in the risk of acute toxicity. In our series, we established a statistically significant correlation between the percentage of bladder volume receiving doses of 60 Gy (V60; p = 0·017), 65 Gy (V65; p = 0·042), 70 Gy (V70; p = 0·012), the prostate gland volume (p = 0.045), and the planning prostate volume (p = 0.029) and the development of acute urinary toxicity. Furthermore, for Pisani et al., Reference Pisani, Galla and Loi10 bladder filling <200cc on CT simulation was predictive of acute toxicity (p = 0.018). Crucially, this observation was not established in our study. Among these initially correlated variables, only the volume of the prostate gland (OR 1·132, 95% CI (1·018−1·260), p = 0·022) and the clinical prostate volume (OR 0·944, 95% CI (0·899−0·991), p = 0·021) are predictive of acute genitourinary toxicity in accordance with our multivariate logistic regression model. The study by Ratnakumaran et al., Reference Ratnakumaran, Hinder and Brand11 on the other hand, demonstrated that the volume of the prostate gland (p < 0·002) was associated with late GU toxicity. His multivariate logistic regression model indicated a predictive value of said volume (OR 1.02, CI (1·00–1·03), p = 0·009) for late GU toxicity.
Regarding late GU toxicity, the analysis of bladder dosimetric parameters from the study by Pisani et al. Reference Pisani, Galla and Loi10 revealed that the maximum bladder dose (greater than 77 Gy) was significantly associated with the appearance of late urinary toxicity (p = 0·013). However, the vast majority of studies have found no relationship between dose-volume parameters and toxicity of GU in the treatment of prostate cancer. Reference Viswanathan, Yorke and Marks4 For Schoeppel et al., Reference Schoeppel, LaVigne and Martel33 no significant difference in late GU toxicity was noted between irradiation doses varying from 68 Gy to 78 Gy. Regarding dose constraints at the bladder level, many authors do not attribute the occurrence of late GU toxicities to the percentage of bladder volume receiving a certain dose. Reference Fuentes-Raspall, Inorizac and Rosello-Serrano34,Reference Storey, Pollack, Zagars, Smith, Antolak and Rosen35 Kuban et al. Reference Kuban, Pollack and Huang36 justify this by the fact that the volume of the bladder changes throughout the treatment. In our study, we did not find any correlations between the occurrence of late GU toxicity and the irradiation parameters and dose-bladder volumes. Nor did we find any predictive value. Like Kuban et al., Reference Kuban, Pollack and Huang36 we believe that this may be correlated to bladder filling, which is not always respected by patients before the dosimetric scan and treatment sessions. But the non-correlation between late GU toxicity and dose-volume parameters in our series could be due to the fact that irradiation did not cause severe late radiation effects. Indeed, late GU toxicities were mainly grade 1, and according to CTCAE v4.0, Grade 1 is mild, not requiring treatment and not significant. Symptoms most often do not require any treatment.
Radiobiological rationale for acute and late genitourinary toxicity
In addition, in radiotherapy, toxicity is partly conditioned by the depletion of functional cells. The clinical manifestation of toxicity results from the destruction of a sufficiently large fraction of clonogenic cells that ensure the renewal of the functional compartment. The intensity of the irradiation delivered during treatment causes the death of a certain quantity of cells, ensuring cell renewal. This principle may explain why our study population developed acute GU toxicities of the order of 50%. An association between dose-volume parameters and acute GU toxicity has thus been established. Spreading the treatment over a sufficiently long period causes the damaged tissues to regenerate compared to the prostate tumour cells. Furthermore, the irradiation of our patients did not cause sufficiently significant long-term destruction of clonogenic cells. It is therefore natural that no correlation between irradiation parameters and late GU toxicity was established in our series.
Limitations of the study on prostate cancer radiotherapy outcomes
However, this study has its limitations. Primarily, the small number of patients involved, particularly for a retrospective study. This is due to a lack of some basic data in patient records, as some patients came from neighbouring countries and returned home after treatment. In addition, another limitation is related to the fact that this study focused exclusively on the cohort of patients who received curative treatment in our institution. Therefore, all patients with distant metastases were excluded. Moreover, given the fact that in our country, the stage of metastatic prostate cancer is predominant on average at 67 years of age (range: 50–79 years) for an estimated population of 1,800,000 inhabitants in 2013, half of whom are under 20 years of age, this considerably reduces the number of cases to be studied.
Conclusion
Our findings underscore the need to establish a robust decision support system to personalise treatment by taking into account not only all relevant dosimetric parameters but also those patient-specific factors that may cause or exacerbate the side effects of radiotherapy for localised and locally advanced prostate cancer.
This work successfully led to the design of two logistic regression models for predicting GU toxicities in localised and locally advanced prostate cancer and, retrospectively, identified significant predictive factors for both acute and late GU toxicities. Specifically, our analysis showed that, acute GU Toxicity was predicted by prostate volume (OR 1·132, CI (1·018–1.260), p = 0·022) and the clinical prostate volume irradiated at the prescribed dose (OR 0·944, CI (0·899–0.991), p = 0·021), late GU Toxicity was predicted by hypertension (OR 7·755, CI (1·105–54.403), p = 0·039) and the development of acute toxicity during radiotherapy (OR 13·243, CI (1.257–139·499), p = 0·032).
We contend that the best way to manage post-radiation side effects is prevention. Consequently, taking into account predictive factors at the origin of genitourinary toxicities linked to radiotherapy treatment of prostate cancer should improve patient-centred decision-making.
This study helps shape future studies aimed at:
-
(i) Identifying specific antihypertensive drugs likely to cause the side effects of radiotherapy by potentially strengthening the action of radiation on tissues;
-
(ii) Developing a formal decision support protocol to facilitate treatment personalisation and
-
(iii) Defining the precise threshold for the proportion of bladder volume receiving doses of 60 Gy, 65 Gy and 70 Gy that causes genitourinary toxicity (GU).
Data availability statement
The research data is stored in our institutional directory and is not accessible to the general public.
Acknowledgements
The authors would like to thank the team of the Department of Medical Physics and Radiotherapy at the Akanda Cancer Institute in Gabon and the laboratory of atomic, molecular and nuclear physics at the University of Yaounde I in Cameroon.
Author Contribution
B.C.M.N. designed the study, collected the data, conducted all statistical tests, analysed, interpreted and reviewed the study and prepared the manuscript. L.G.J. designed the study, collected the data, and analysed, interpreted and reviewed the study. S.Y.L.M. designed the study and analysed, interpreted and reviewed the study. P.O.M. designed the study and analysed, interpreted and reviewed the study. A.B.K., E.B. and G.H.B.-B. supervised the study. All authors have read and approved the manuscript.
Financial support
This work has received operational support from Akanda Cancer Institute, Gabon.
Competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Ethical approval
In accordance with the ethical principles of the Declaration of Helsinki, all patient data were anonymised before assessment. The Institutional Review Boards of the Akanda Cancer Institute ruled that, given the anonymised and retrospective nature of the study, formal approval was not required.
Consent for Publication
All authors read and approved the final manuscript version.
Open access
