A variety of tissue preparation strategies have been used to generate PDEs. The most common slicing methods are manual dissection with a scalpel and mechanized sectioning using either the vibratome (a vibrating blade microtome) or the McIlwain Tissue Chopper. Manual slicing is used to dissect excess fat or generate PDEs within a specific size range (30), however, it is becoming less common due to the lack of uniformity in explant thickness (16, 22). Two studies have directly compared the most common mechanized slicing technologies, the vibratome and Tissue Chopper, and found the Tissue Chopper more accurately and reproducibly produced slices of a specified thickness (19, 25). Furthermore, Holliday, Moss (21) were unable to slice breast tissue using a microtome due to the fatty nature of the tissue. By contrast, breast explants were successfully sliced using the Tissue Chopper resulting in improved tissue integrity when compared to a vibratome (22). Thus, where available, the McIlwain Tissue Chopper is currently the preferred approach for accurate slicing of tissues of varying densities.

Supplements incorporated in PDE culture media are highly dependent on the tissue of origin (summarized in Table 1 ). When comparing the studies that used tissue specific supplements, there was no clear conclusion as to whether they improved longevity (22, 26, 31). However, Naipal, Verkaik (22) directly compared four types of media, two breast cancer-specific and two common culture media combinations, and found that the tissue specific media (listed in Table 1 ) was most effective. For HNSCC, Lee, You (26) reported that PDEs remained viable for 10 days, longer than the 2-6 days reported for other HNSCC PDE cultures in more basic media (25, 38). By contrast, in breast and pancreatic PDE studies the addition of tissue specific supplements did not improve viability compared to less complex culture media formulations (22, 31).

It is apparent that a combination of sodium bicarbonate, HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) and cysteine has a positive effect on viability of PDE cultures across several tumor types. Gerlach, Merz (25) and Jiang, Seo (8) used these supplements in HNSCC and pancreatic cancer studies, respectively, where both had increased longevity compared to other studies. Sodium bicarbonate and HEPES are pH buffering reagents used in culture to control acidity (39). Cysteine is an essential amino acid important for protein synthesis and other metabolic functions, and helps promote proliferation in cancer cells (40). Zhang, van Weerden (41) found that a tissue-specific composition and media containing sodium bicarbonate helped improve viability compared to non-tissue-specific formulations when using PDX material. Given the fundamental roles of these supplements, they are likely to also be relevant to improving the longevity of explants for multiple tumor types, although this remains to be tested.

Several culturing techniques have been adopted across tumor streams for PDE culture. The three most commonly used techniques are free-floating culture, grid or pore membrane supports, and gelatin sponge supports (42). Additionally, specialized platforms such as hydrogels and co-culture systems have been investigated in improving longevity of PDE cultures (26, 43, 44).

Free-floating culture of PDEs in the absence of rotation has been largely dismissed due to the overgrowth of cells resulting in monolayer growth (42), increase in apoptosis, and loss of tissue integrity (41). However, incorporating orbital shaking into this approach reduces these issues. Ovarian PDEs were still viable, proliferating, and metabolically active at the end of a 30-day orbital shaking, free-floating culture period, and no necrotic core was present indicating successful diffusion of nutrients (17). Similarly, Naipal, Verkaik (22) adopted this technique for breast cancer PDEs, suggesting it was most effective at nutrient diffusion and easy to replicate. By contrast, Davies, Dong (28) compared free-floating under rotation to a pore membrane in breast and prostate PDX cultures finding that the membrane culture was more successful at improving viability. Histological examination, and immunohistochemistry (IHC) for E-cadherin were performed in this study, where floating cultures showed the presence of more vacuoles, loss of E-cadherin, reduced proliferative capacity, and increased apoptosis compared to the membrane cultures, confirming that membrane cultures were able to maintain structural integrity over a longer culture period (28).

The use of a grid or pore membrane has been most adopted (19, 28). The most common membrane is the Millipore system due to its open access system allowing for easy transfer and access to culture media, enabling efficient gas exchange between the PDEs and culture media (25, 27, 31). Millipore inserts coated in collagen matrix derived from rat tail have also been used to culture pancreatic PDEs (8, 15). Lim, Chang (15) commented that the precoated membrane increased adherence of the tissue to the surface. However, there was no significant increase in viability using the precoated membrane compared to standard Millipore inserts.

The use of gelatin sponges that support the PDE at the air-liquid interface is a relatively new culture technique. Centenera, Raj (42) noted that sponges help the exchange of nutrients and waste by acting as a capillary surrogate in prostate PDE cultures. Centenera, Hickey (20) and Kokkinos, Sharbeen (16) also used sponges for PDE culturing, demonstrating a significant increase in viability compared to the other culture techniques, such as filter and membrane cultures in prostate and pancreatic PDE studies, respectively. The use of sponges aids in avoiding overgrowth of a cell monolayer from cells disseminating from the explant and helps maintain homeostasis through the exchange of nutrients and waste (20).

Hydrogels that mimic features of the extracellular matrix in tissue culture have become increasingly popular, due to progression from 2D to 3D culture platforms. A preliminary study by Hribar, Wheeler (43) investigated the feasibility of embedding PDEs in VersaGel^®, a growth factor-free hydrogel. After three days tumor cells moved out of hydrogel-embedded explants, which was associated with remodeling of the extracellular matrix and mimicking tumor invasion. An advantage of hydrogels is the ability to retain secreted growth factors and to incorporate tissue-specific growth factors into the scaffold (45). These advantages allow for a highly customizable functionalized platform for PDE growth that can be used across tumor types.

Specialized experiments have been performed requiring unique culturing techniques that have not been widely adopted due to complexity and/or technical challenges. Given that angiogenesis is a hallmark of cancer and contributes to the complexity of the TME, this is a critical component missing in routine PDE cultures. Bazou, Maimon (44) used vasculature beds formed from co-culturing endothelial and smooth muscle cells to mimic angiogenetic formation in pancreatic cancer, resulting in PDE viability extending longer than most other culture systems at 3 weeks. Similarly, Lee, You (26) positioned HNSCC PDEs on a cell sheet comprised of matched patient epithelial and sub-epithelial cells to create a co-culture system. The results showed improved cell viability and longevity of PDEs when comparing the co-culture system to other non-matrix models and showed a reduction of viable cancer cells following SOC treatments. Despite the promise of these unique culture techniques, they are not readily available, difficult to establish, and expensive, making them a less attractive technique for widespread use.

There is limited data addressing different cell culture incubation conditions, with most studies using 37°C, 5% CO₂ and 21% oxygen. There is no published data addressing variation of temperature or CO₂ concentration, however the effect of different oxygen levels on PDE viability has been examined. Misra, Moro (31) compared hyperoxic (41%), normoxic (21%) and hypoxic (less than 21%) oxygen levels concluding that hypoxic oxygen levels led to tissue degradation, but there was no significant difference in viability, proliferation, hypoxia, and metabolic activity between normoxic and hyperoxic conditions. Additionally, Davies, Dong (28) showed that lung and prostate PDX explants maintained in atmospheric oxygen (21%) had improved viability compared to tissues cultured under low oxygen conditions. Thus, atmospheric oxygen is likely to be most suitable for PDE incubation.

PDEs have been widely used for the evaluation of tumor response to chemotherapeutics, modeling clinical treatments that patients may receive as SOC (30, 56, 60). The possibility of resistance to chemotherapies means it is useful to test multiple regimens, where they are available, to determine the treatment(s) most likely to be effective for the individual patient. Merz, Gaunitz (19) treated glioblastoma PDEs derived from different tumors with temozolomide, a first line SOC agent, for 3 days, and found that individual tumors displayed different susceptibility to the treatment, consistent with clinical observations. Additionally, Ricciardelli, Lokman (30) treated PDEs with SOC chemotherapies such as carboplatin, which induced apoptosis in PDEs sensitive to chemotherapies but not in chemoresistant PDEs. This reinforces the relevance of adopting pipelines to probe patient-proximal models prior to clinical therapies.

Immunotherapies have increasingly become first-line for advanced cancers such as melanoma (61), colorectal (62) and lung (63, 64), due to outstanding tumor control in a subset of patients and better tolerability when compared to chemotherapy (65). While in its infancy, a growing number of tumor types, including endometrial (57), colorectal (54, 55) and head and neck (24, 38) cancer have adopted ex vivo treatment of PDEs to study CTL recruitment and immunotherapy efficacy. Seo, Jiang (66) showed that pancreatic ductal adenocarcinoma cell killing within the slice was mediated by CTL activity following 6-days of combined PD-1 and C-X-C chemokine receptor type 4 (CXCR4) blockade. Interestingly, anti-PD-1 (nivolumab) treatment of PD-L1⁺ melanoma PDEs increased the inter-cell distance between T cell subsets, demonstrating mechanisms whereby CTLs may avoid Treg-mediated immunosuppression in the TME following immunotherapy (67). As the numbers of studies incorporating PDEs with immunotherapy grows, future pipelines could be assessed in parallel with multiple mRNA and protein readouts to assess immunological response in patients eligible for anti-PD-1, anti-PD-L1 or anti-CTLA-4 treatment, as described in the Lombard Street Approach (68).

Therapies focusing on exploiting specific molecular components to destroy proliferating cancer cells (69) are an important component of successful personalized therapeutics testing using PDEs. Although the concept of targeted therapies is more attractive than non-specific systemic treatments, side effects and resistance remain issues (70, 71). In that regard, targeted therapies are ideally informed by a WES pipeline to identify suitable and relevant candidates. Any molecular targets identified can be compared to FDA approved target therapies using ex vivo testing. Currently the feasibility of incorporating precision target therapies with functional testing of individual patient samples is poor due to the misalignment of WES turnaround times (48, 52, 72) and short-term viability associated with PDEs ( Table 1 ). Nonetheless, targeted therapies such as erlotinib (EGFR inhibitor) and BYL719 (PI3K inhibitor) were shown to suppress the mitogen-activated protein kinase (MAPK) and protein kinase B (AKT) pathways respectively in PDEs (38). These signaling pathways are commonly upregulated during the tumorigenesis process, stimulating increased proliferation of cancer cells (73). PDE platform testing approaches can be used to determine whether such therapies might be useful for individual patients and avoid using treatments to which tumor cells are resistant and thus unnecessarily toxic.

The clinical predictive value of PDE culture as a patient-proximal testing platform needs further investigation. Brijwani, Jain (54) and Majumder, Baraneedharan (74) have used an ex vivo platform, based on PDE drug response, genomic and machine learning (CANscript), to predict patient responses in locally advanced/metastatic CRC and HNSCC. Their data reports that the sensitivity and specificity of this platform for clinical prediction exceeds 90% (54, 74).