Enhancing Prostate Tumor Biobanking Reliability with Improved Sampling Technique and Histological Characterization

Prostate cancer (PC) is the 2^nd most frequent cancer in men and the 5^th leading cause of death worldwide¹. Patient treatment and prognosis depend on the staging and grading (Gleason score) of the tumor, as evidenced by the higher 5-year survival rates of localized and low-grade tumors (Gleason grade 6) (99%) compared to high Gleason grades and metastatic tumors (31%)².

PC local relapse and treatment failure have been linked to the characteristic high genetic intratumor heterogeneity of this tumor type³. Additionally, PC is considered to be a multifocal disease with several tumor foci exhibiting different morphological, histological, and molecular characteristics⁴, which may originate independently or derive from a common tumor cell ancestor⁵. Previous studies have shown that tumor evolution differs among patients based on specific genetic drivers that can promote metastasis or confine the cell lineage to the prostate⁵. Therefore, molecular characterization of the different tumor foci is crucial not only for providing a more accurate diagnosis and prognosis but also for tailoring effective and personalized treatment for the patient.

In this context, biomedical research and integrative multi-omics approaches are offering unprecedented opportunities to classify cancers into different subtypes, identify diagnostic and prognostic biomarkers, and discover markers related to treatment response. Furthermore, these approaches contribute to a better understanding of the biology of this disease^6,7. Biological samples, whether tissues or biofluids, can be analyzed using various multi-omics platforms (genomics, transcriptomics, proteomics, metabolomics, etc.) to uncover the biological features underlying cancer pathophysiology, thereby addressing current limitations related to genetic and phenotypic heterogeneity⁶. However, it's important to consider that the quality of data derived from omics studies depends on the quality of the samples collected from tumors, their accurate characterization, and subsequent processing and storage⁸.

In this context, obtaining fresh PC tissue for research presents a methodological challenge due to the difficulty of successful tumor sampling⁹. Previous methods involved random sampling following radical prostatectomy, yielding poor results¹⁰. However, more recent approaches incorporate targeted protocols based on both magnetic resonance imaging (MRI) and biopsy data, resulting in improved efficacy in tumor sample collection¹¹.

On the other hand, histopathological characterization of samples prior to their storage without significant tissue alteration also poses an interesting challenge. Consequently, in many cases, the histopathological determination of samples is performed after their analysis (e.g., HR ¹H NMR metabolomic analysis)¹². This practice entails unnecessary expenses, time consumption, and the loss of a significant number of samples that are eventually excluded from the analysis (for example, samples that, following histopathological analysis, turn out not to be tumor samples). In other cases, the histopathological characterization of samples is performed before their analysis. In fact, some previous studies have attempted to standardize methods for providing representative high-quality research samples from radical prostatectomy specimens for genomics and metabolomics^13,14. Nevertheless, sampling efficiency is significantly higher when performed from already histologically confirmed sections (88%) that disrupt tissue, compared to when performed from unconfirmed sections (45%)¹.

Here, a new methodology is presented to overcome these limitations, aiming to obtain fresh and well-characterized PC samples before storage in the Biobank. This method has been developed through collaborative efforts between different clinical services (Urology, Pathology, and the La Fe Hospital Biobank). It's important to highlight that Biobanks play an essential role in the collection, processing, preservation, and storage of biological samples while ensuring the high quality of samples and data, as well as compliance with ethical and legal requirements^8,15,16.

This method was developed through collaborative efforts involving different clinical services (Urology, Pathology, and the La Fe Hospital Biobank). The study was conducted in compliance with institutional, national, and international guidelines for human welfare, and it received approval from the Ethics Committee for Biomedical Research at the Instituto de Investigación Sanitaria Hospital Universitario y Politécnico La Fe (Valencia, Spain). All samples were stored at the La Fe Hospital Biobank (PT13/0010/0026). The overall procedure is detailed in Figure 1.

Figure 1: Overall procedure scheme. Schematic representation of the stepwise procedure described in the protocol section. Please click here to view a larger version of this figure.

1. Tumor targeting

1. Before surgery, utilize multiparametric or biparametric Magnetic Resonance Imaging (mpMRI or bpMRI)^17,18 to locate the tumor and map its position within the prostate gland. Combine the information obtained from MRI with findings from transrectal ultrasound or rectal examination if the latter yields discordant results.
2. To enhance tumor localization, review the results of preoperative ultrasound-guided and MRI-guided transrectal prostate biopsies.
   ​NOTE: In the diagnostic procedure for prostate cancer, biopsies are taken directly from the prostate via transrectal access. These biopsies include systematic sampling from the prostate (involving cylinders distributed all around the gland) and cylinders directly targeted to suspicious findings on MRI. To target the retrieval of these samples, transrectal ultrasound is used, which is combined with MRI findings in a cognitive manner ("cognitive fusion") or with "real MRI-ultrasound fusion software" when available.

2. Radical prostatectomy and prostate collection

1. Perform a standardized robotic-assisted radical prostatectomy through a transperitoneal approach, following the steps outlined by Huynh et al.¹⁹.
   1. Start with abdominal access, insufflation, and camera port placement (see Table of Materials). Place the remaining five ports under direct vision. Set pneumoperitoneum to 12 mm.
   2. Release the bladder by incising the peritoneum and dissect the perirectal space.
   3. Open the pelvic fascia on each side.
   4. Perform a complete dissection of the anterior and lateral aspects of the prostate and bladder.
   5. Carry out the opening of the bladder neck, identify and section the proximal urethra to gain access to the posterior aspects of the prostate and bladder.
   6. Dissect the seminal vesicles and rectum space¹⁹.
   7. Dissect the lateral aspects of the prostate, sparing the neurovascular bundles.
   8. Ensure hemostasis of vessels in the retropubic space.
   9. Dissect the apex and section the distal urethra.
   10. Complete the vesicourethral anastomosis.
   11. Extract the prostate and perform the closure of the incision¹⁹.
2. Collect the fresh prostate in a dry container. It is crucial not to fix or paraffin-embed the sample specimen.
3. Transport the fresh surgical specimen, which includes the entire prostate, to the Biobank for sample processing.

3. Prostate processing and sample collection

1. Select targeted prostate zones for sampling based on tumor location information retrieved from preoperative mpMRI, preoperative biopsy results, and direct prostatectomy specimen palpation. Record this information on a form that includes patient identification data, prostate areas finally sampled, and the number of samples collected.
   NOTE: When MRI findings, preoperative biopsy results, and direct prostatectomy specimen palpation are concordant, target all the samples to that or those locations (between 4 and 6 cylinders). If there is any discordance, increase the number of cylinders and separately target all zones that are suspicious of being affected by any of those methods.
2. Use a disposable automatic needle (surgeon) to retrieve cylindrical samples from the selected areas (step 1.1). Then, identify (pathologist) the outer zone of the cylinder, clamp the cylinder with tissue forceps, place it on a slide, and mark the outer zone with tissue ink. Repeat this process with all cylinders taken.
3. Place the collected cylinders horizontally on a sample holder disk (see Table of Materials) filled with cryostat embedding medium and transfer them to the microtome cooling platen for cutting.
4. Once placed on the platen, attach the sample holder disk to the specimen chuck, and cut it with careful movements of the wheel following the next steps:
   1. Use the optimal cutting temperature compound (OCT compound, see Table of Materials) to embed the tissue sample before frozen sectioning on a microtome-cryostat.
   2. Afterward, cut the sample using a cryostat-microtome at -25 °C. Collect thin cylinder-shaped sections of tissue (10 µm) onto treated microscope slides. Ensure to collect only a few microns of tissue to ensure an adequate amount of intact tumor tissue for storage in the Biobank.
   3. Gather two or three thin sections per slide and manually stain them to control the final coloration. The technician's eye ensures optimal staining, and any necessary re-staining can be done. Follow these steps
      1. Fix slides in 96% ethanol, hydrate in distilled water, and stain with Harris hematoxylin (see Table of Materials) for 1 min to stain the nuclei. Remove excess hematoxylin by rinsing the slides with distilled water.
      2. Treat slides with 30% ammonium hydroxide to turn the hematoxylin blue. Remove excess solution with rinses of distilled water.
      3. Stain slides with eosin (see Table of Materials) for 30 s.
      4. Next, dehydrate the slides with increasing titrations of alcohol, clear with xylene (100%), and mount them with DPX (a mixture of distyrene, a plasticizer, and xylene, used as a synthetic resin mounting media that replaces xylene-balsam, see Table of Materials). During these steps, dip the slides into and remove them from the jar of Coplin solution.
      5. Finally, place a coverslip on top of the stained tissue and use a medium with a suitable refractive index (see Table of Materials) for viewing under a microscope.

4. Sample characterization

1. Analyze the sample under a light microscope (pathologist), discriminating the areas of carcinoma and determining the Gleason score of the tumor. Document all these data and mark the precise tumor zones on the slides.
2. During this step, label the slide with the precise location of the tumor and its respective Gleason Score²⁰. Also, mark the location of PIN (Prostatic Intraepithelial Neoplasm) based on morphological identification.

5. Sample micro-macro-dissection and storage

1. Clearly mark different histological components with distinct markers on the slide (pathologist) and directly correlate them with frozen tissue on the disk.
2. Then, using a surgical blade, macro-micro dissect the areas of the sample marked on the slide within the OCT disc, dividing them into tumoral, peri-tumoral, PIN, and normal prostatic tissue.
3. As the embedding medium begins to defrost, handle the tissue with forceps and place it into correctly labeled cryotubes.
   NOTE: Carefully store each dissected fragment within the corresponding tube, following the histological diagnosis of the cryo-slices.
4. Store the labeled cryotubes with fully characterized histological samples in the Biobank at -80 °C for subsequent use in research studies.
   NOTE: The steps described in this section can be performed by a trained technician under the supervision of the pathologist. The surgical specimen should be transported to the Biobank as soon as possible to prevent tissue and nucleic acid degradation. According to the reference guides, efforts must be made to avoid prolonged ischemia time (>30 min)²¹. The samples can be kept in the Biobank indefinitely as long as the sample quality principles are maintained.

The results reveal that this technique has made it possible to obtain tumor material in 61% of the cases studied (25 out of 41 cases) (Table 1).

Table 1: Histopathological data of study samples. Summary of histopathological data for the samples used in the study. The diagnostic cylinder refers to the prostate biopsy sample obtained for diagnostic purposes, while the processed cylinder corresponds to the cylinder obtained from the prostatectomy specimen for research purposes. ND: no data; LL: left lobule; RL: right lobule; TTB: total tumor burden. Please click here to download this Table.

For this analysis, the number of samples with the presence of tumor obtained in the cylinder acquired from the radical prostatectomy specimen from each of the lobules was compared with the presence of tumor in the cylinders of the biopsy performed for diagnosis. Specifically, the data reveal an efficiency of 59% (17/29) in the biopsies performed on the left lobe and 21% (7/33) in those performed on the right lobe (Table 1). Regarding the 16 prostates from which it was not possible to obtain a tumor during the sampling, 10 of them had a tumor load of less than 10%, and in no case did any of them present a tumor load of more than 20% (Table 1). A Fisher correlation test showed a capacity for obtaining statistically significant tissue when the tumor volume is at least 12% (p = 0.0187).

Finally, a receiver operating characteristic (ROC) curve analysis was carried out to determine the predictive performance of this method, obtaining an area under the curve (AUC) of 0.843 (Figure 2). These data indicate that this newly-developed method is capable of delivering satisfactory results with regard to tumor acquisition.

Figure 2: ROC curve. Receiver Operating Characteristic (ROC) curve illustrating the model's performance in binary classification of tissue samples as either tumor or non-tumor tissue. Please click here to view a larger version of this figure.

Lastly, to evaluate the concordance between the Gleason score of the tumor cylinder obtained by sampling and the Gleason score of the radical prostatectomy specimen, a Pearson's chi-squared analysis was performed. The results provided a concordance rate of 64%, increasing to 96% when immediate superior or nearest superior Gleasons were included (i.e., 3+ if 3+4; and 3+4 if 4+3). Figure 3 shows an example of the histopathological correlation of the radical prostatectomy specimen vs. the processed cylinder biopsy. It is important to note that all different histological patterns can be found in a single tissue sample: a tumor pattern with Gleason 3, tumor Gleason 4, PIN, and/or normal tissue.

Figure 3: Histopathological correlation of radical prostatectomy specimen vs. processed cylinder biopsy. (A) Hematoxylin and eosin (H&E) staining of needle core biopsy at 20x magnification, depicting Gleason pattern 3. Small, well-formed glands resembling normal prostate tissue are observed, lined by cancerous cells with distinctive features. (B) H&E staining of the specimen at 40x magnification, Gleason pattern 3. (C) H&E staining of needle core biopsy at 20x magnification, illustrating Gleason pattern 4. Irregularly shaped glands with enlarged, atypical cells are evident. (D) H&E staining of the specimen at 40x magnification, Gleason pattern 4. (E) H&E staining of needle core biopsy at 20x magnification, demonstrating Prostatic Intraepithelial Neoplasia (PIN). Crowded, overlapping cells within ductal spaces are observed. (F) H&E staining of needle core biopsy at 40x magnification, depicting PIN. (G) H&E staining of needle core biopsy at 20x magnification, representing normal prostate tissue. (H) H&E staining of needle core biopsy at 40x magnification, normal tissue. Scale bars = A,C,E,G,H, 0.5 mm; B,D,F, 0.05 mm. Please click here to view a larger version of this figure.

In conclusion, the most representative outcomes of the method presented herein include, on the one hand, the ability to obtain fresh intact samples (not fixed or paraffinized) characterized at the histological level to develop quality omics studies (genomic, transcriptomic, proteomic, metabolomic, among others). Additionally, it reduces bias associated with tumor heterogeneity in research studies, because each piece of micro- and macro-dissected tissue, stored in individual cryotubes, has a unique histology (Gleason, preneoplastic, normal tissue). The microscopic observation of the sample obtained makes it possible to define the different histological patterns found along the cylinder (Figure 3). Each tissue region with a given histological pattern is individually referenced, cut, and stored in a cryotube. Thus, each cryotube contains a tissue sample with unique histology. The analyses carried out on each of these samples will allow one to overcome the drawbacks of tumor heterogeneity. Other representative benefits include the reduced cold ischemia time in sample preprocessing. The average time from extraction to inclusion of the tissue sample in OCT was 13 min, and to storage, it was 33 min. This fact ensures better DNA and RNA quality. Finally, the establishment of fully characterized collections with information on clinical and pathological data, plus images of each of the samples obtained, guarded by a Biobank, can be used by the scientific community worldwide.

In any research study, obtaining quality samples is an essential requirement to reduce systematic biases and obtain reliable results²². Therefore, control of preanalytical variables such as the temperature at which samples are processed and stored, the time elapsed from sample collection to storage, the use of sterilized materials, or the effects that the addition of preservatives or other additives can have on samples must be considered in any protocol involving biological samples. Not only is this key, but also the correct identification and codification of samples, as well as their pathological characterization, are crucial. Otherwise, there might be a large number of high-quality samples stored but remaining unused.

The collection and typification of tumor and non-tumor samples prior to storage in the Biobank are the main focus of this protocol due to their vast importance in the context of PC. Obtaining fresh PC tissue for research poses quite a challenge from a methodological point of view, due to the difficulty of successful tumor sampling⁹. Usually, targeted sampling in the prostate to collect tumor samples for research is not possible. Another limitation includes obtaining a representative tumor sample of the surgical specimen. The method presented here has successfully overcome these limitations, showing good performance in the acquisition of tumor samples, especially when the tumor burden exceeds 12% (which happens in most cases). Additionally, this method has provided a Gleason score correspondence between tissue cylinder and radical prostatectomy of 64%, rising to 96% when immediate superior or nearest superior Gleasons are included. These results are more than satisfactory, especially when compared to other studies previously performed for diagnostic purposes that show lower concordances^23,24,25.

Moreover, compared to other published protocols, those with better results always require manipulation of the prostate specimen and a delay in the study. But correlating mpMRI, diagnostic biopsies, and manual inspection of the specimen prior to sample acquisition ensures less manipulation and better results than random sampling.

On the other hand, when a biopsy is collected, the tissue obtained can contain a heterogeneous population of cells with different histological characteristics, i.e., normal tissue, tumoral tissue with a low Gleason score, tumoral tissue with a high Gleason score, PIN, atypical small acinar proliferation (ASAP), etc. Consequently, the characterization and micro-macro dissection of this sample before carrying out the analysis allows for a more precise selection of the samples of interest and leads to more robust studies. Therefore, the greatest value of this methodological protocol lies within its ability to sample and characterize prostate tumor samples, ensuring high-quality samples and a complete report of their histological characteristics prior to storage. Although in this case all radical prostatectomies have been performed using the minimally invasive robotic surgical system, obtaining the sample using conventional techniques does not represent a problem. However, some critical steps have to be considered to obtain good results. On the one hand, control of the cold ischemia time between sample collection and processing, as this is directly related to DNA, RNA, and protein quality and integrity²⁶. For this, the working team must be truly involved and implicated. On the other hand, it is crucial to have a pathologist or a technician with expertise in prostate cancer to typify, straightaway, the sample obtained to perform the macro-micro dissection of the sample.

Regarding the limitations of the method, the need to obtain a minimum amount of tissue from the prostate piece by means of targeted punctures guided by previous clinical data (i.e., MRI and biopsy data) can be highlighted. Another drawback is that this protocol can only be applied to a subgroup of prostate cancer patients, those undergoing radical prostatectomy. Finally, the use of OCT to embed tissue samples prior to frozen section on a microtome-cryostat may also be considered an impediment, especially in the analysis of data derived from metabolomic studies. However, this fact can be overcome by a series of washing steps, so the use of OCT is compatible with metabolomic studies as demonstrated in previous studies^27,28,29.

In conclusion, the protocol presented here has potential applications in the field of prostate cancer research, but it can also serve as an insight for the collection and characterization of other types of human samples. Additionally, this protocol overcomes current limitations related to tumor heterogeneity, which has a great impact on the development of "omics" studies applied to biomedical research. These studies may further the identification of diagnostic, prognostic, or treatment response biomarkers, leading to the detection of subgroups of patients based on molecular characteristics.