2.1 Ethical Statement
This study was approved by the University of Pisa Medical School Ethical Committee (ID # 19005), and it has therefore been performed in accordance with the principles embodied in the World Medical Association Declaration of Helsinki. All participants gave their informed consent prior to their inclusion in the study.
2.2 Participants and Procedure
Participants were recruited among patients with four types of urogenital cancer: testicular cancer (TC, n=15), prostate cancer (PC, n=69), renal cancer (RC, n=34) and bladder cancer (BC, n=19) treated at the Departments of Surgical, Medical, Molecular Pathology and Critical Area and Department of Translational Research and Advanced Technologies in Medicine, University of Pisa. In total, one hundred-thirty-seven patients were recruited (mean age = 66.9±14.9 SD years) as cancer patients and another 21 volunteers matched in age and not affected by cancer as a comparative group (mean age = 67.3±8.8 SD years).
Participants were tested on a week-day morning between 9:00 am and 12:00 pm to minimize circadian variations of steroid levels. Because the collection time was at least 2 h after awakening, cortisol awakening response was not an issue. On arrival, participants initially rested for a few minutes while they were informed of the procedure and of the general goals of the research. After they gave their informed consent, the questionnaires were administered (the average completion time was less than 30 minutes per patient). After that, saliva samples were collected by passive drooling. Participants were given the instructions to hold a short straw in their mouth with the test tube at the end and let the saliva drool into the test tube. Samples were stored at -70°C within 30 min of their collection.
2.3 Self-report Assessments
The Functional Assessment of Cancer Therapy-General (FACT-G) was designed to measure four domains of quality of life in cancer patients: physical, social, emotional, and functional well-being [19]. The FACT-G comprises four subscales: Physical Well-Being (PWB), Social/Family Well-Being (SWB), Emotional Well-Being (EWB), and Functional Well-Being (FWB). From the data we were able to generate an overall score with range and distribution specific to our sample.
The Profile of Mood States (POMS) [20] and the Personality Belief Questionnaire – Short Form (PBQ–SF) [21] were developed as clinical and research instruments to assess dysfunctional emotional regulation and beliefs associated with individual personality disorders. When used in cancer research, these assessments were mostly used in relation to the ability of patients to deal with negative biopsy results [22]. The POMS questionnaire was employed to assess current mood states and mood changes in our sample [23]. These scales have been found to be consistent and reliable in both clinical and non-clinical samples, including patients with cancer [24] and immune dysfunctions [25]. The PBQ-SF questionnaire was used to assess potential sets of dysfunctional beliefs associated with cancer in our sample [26]. According to current cognitive theories, each personality disorder is characterized by a specific set of dysfunctional beliefs [26].
The Life Events Checklist for DSM-5 (LEC-5) is a comprehensive screening instrument used to detect exposure to a range of potentially traumatic, life changing events [27-28]. Despite its widespread use in studies focused on posttraumatic stress symptoms and disorders [29], this instrument has not been used as much to assess psychosocial distress among oncology patients, even though a large proportion of people diagnosed with cancer experience levels of distress that would benefit from psychosocial interventions [30].
2.4 Endocrinological Assessments
Prior to the assays, saliva samples were shaken vigorously in a mixer for approximately 30 s. Next, the tube was centrifuged for 15 min at 1250g. Using a transfer pipette, the supernatant was transferred to a 13x100mm glass test tube. This process was repeated three times to assure no contaminations were introduced in the final steps. The final steps of the assay procedure were performed according to the instructions of the manufacturer of the assay kit (Salimetrics, State College, PA, USA). Sample readings were completed using an automated micro-plate reader (BioTek Instruments, Winooski, VT, USA; model: Hybrid), and the Gen5 software (BioTek, BioTek Instruments, Winooski, VT, USA; version 2.01). Readings were assessed at a wavelength of 450 nm, with secondary filter corrections at 490 to 492 nm.
Quality controls were determined using 20 replicates at five different concentrations for each hormone. Intra-assay precision showed an average coefficient of variation of 3.9% for DHEA, 4.2% for DHEA-S, and 3.5% for CORT. The inter-assay precision was calculated in a similar way and returned the following percentages: 4.5% for DHEA, 6.1% for DHEA-S, and 3.8% for CORT. To calculate recovery rate, five samples containing different levels of an endogenous hormone were spiked with known quantities of the same hormone and assayed. Average recovery ranges from 95% to 110% for all three hormones. The functional sensitivity of the kits was as follow: 8.32 pg/mL for DHEA, 198.3 pg/mL for DHEA-S, and 0.018 µg/dL for CORT.
2.5 Statistical Analysis
Analysis of variance (ANOVA) and t-test were used to analyze the difference in the averages by cancer type and presence or absence of metastases. Correlation among variables was assessed using Pearson’s product-moment coefficients. Analysis of covariance (ANCOVA) was used to control for the age if the participants. Although using many different tests to extract information from the same database can increase the probability to Type II errors, we used these many analyses as general guidance for the multivariate test explained below. Nevertheless, significance levels between 0.05 and 0.01 should be taken cautiously. All tests were two-tailed, and the significance threshold was set at α=0.05.
To evaluate the internal consistency of all self-reported measures, we calculated the Cronbach’s a. This measure is computed by correlating the score for each item with the total score for each observation, and then comparing it the variance for all individual scores. The resulting coefficient ranges from 0 (all the items are independent from each other) to 1 (all the items have high covariance). Generally, coefficients higher than 0.60 are considered acceptable and higher than 0.80 are considered representing high internal consistency. All four self-reported measures in our dataset returned acceptable internal consistency, ranging from 0.67 (LEC-5) to 0.89 (FACT-G).
Logistic regression analysis was used to identify the best predictors of the binary dependent variable presence (1) or absence (0) of metastases. We selected this approach to provide a multivariate model able to summarize the many univariate analyses presented previously, thus reducing the probability of an inflated significance estimation. Predictors were DHEA, DHEA-S, CORT, age, and the self-report measures. The Wald statistic was used to assess the significance of each coefficient. The overall goodness-of-fit was assessed by the -2 Log Likelihood and by the Hosmer and Lemeshow test. The percentage of variance explained by each model was assessed using the Cox and Snell R2 and the Nagelkerke R2. The predictive accuracy of the models was determined by the hit ratio, that is the percentage of cases correctly classified.
All statistical analyses were conducted using the SPSS computer program (IBM, Chicago, IL, version 27.0).