Personalized Identification of Optimal HIPEC Perfusion Protocol in Patient-Derived Tumor Organoid Platform

Abstract

Background.

Chemotherapy dosing duration and perfusion temperature vary significantly in HIPEC protocols. This study investigates patient-derived tumor organoids as a platform to identify the most efficacious perfusion protocol in a personalized approach.

Patients and Methods.

Peritoneal tumor tissue from 15 appendiceal and 8 colon cancer patients who underwent CRS/HIPEC were used for personalized organoid development. Organoids were perfused in parallel at 37 and 42 °C with low- and high-dose oxaliplatin (200 mg/m² over 2 h vs. 460 mg/m² over 30 min) and MMC (40 mg/3L over 2 h). Viability assays were performed and pooled for statistical analysis.

Results.

An adequate organoid number was generated for 75% (6/8) of colon and 73% (11/15) of appendiceal patients. All 42 °C treatments displayed lower viability than 37 °C treatments. On pooled analysis, MMC and 200 mg/m² oxaliplatin displayed no treatment difference for either appendiceal or colon organoids (19% vs. 25%, p = 0.22 and 27% vs. 31%, p = 0.55, respectively), whereas heated MMC was superior to 460 mg/m² oxaliplatin in both primaries (19% vs. 54%, p < 0.001 and 27% vs. 53%, p = 0.002, respectively). In both appendiceal and colon tumor organoids, heated 200 mg/m² oxaliplatin displayed increased cytotoxicity as compared with 460 mg/m² oxaliplatin (25% vs. 54%, p < 0.001 and 31% vs. 53%, p = 0.008, respectively).

Conclusions.

Organoids treated with MMC or 200 mg/m² heated oxaliplatin for 2 h displayed increased susceptibility in comparison with 30-min 460 mg/m² oxaliplatin. Optimal perfusion protocol varies among patients, and organoid technology may offer a platform for tailoring HIPEC conditions to the individual patient level.

Conducting prospective randomized clinical trials for peritoneal malignancies has been historically challenging, resulting in variability in perfusion protocols between institutions. Mitomycin (MMC) and oxaliplatin are the most commonly perfused chemotherapy agents. The evidence comparing the efficacy, optimal dose, perfusion length, and impact of hyperthermia and whether the benefits are generalizable for all the patients is still a subject of debate.^1–7 Currently, decisions on hyperthermic intraperitoneal chemotherapy (HIPEC) protocols are based primarily upon pharmacokinetics. Pharmacokinetics do not capture residual postperfusion tumor cell viability, achieved by a specific drug concentration. In a similar fashion, it is not possible to quantify the effect of intraperitoneal hyperthermia based on the type of primary tumor or the type of chemotherapy at the level of the individual patient; nor to potentially identify patients in whom the tumor possesses the biologic machinery to resist heat.

Growing evidence utilizes patient-derived tumor organoids (PTO) as a therapeutic response prediction platform.^8–11 Reconstructing the patient’s tumor in the forms of patient-specific organoids has several advantages over conventional cell cultures and patient-derived xeno-graft (PDX) models. This includes much higher take rates and three-dimensional (3D) recapitulation of tumor microenvironment, including not only tumor cells but also associated stroma and incorporated immune system as well.^12–14 Results can be obtained within 10 days from tissue biopsy, which represents a timeframe aligned with the patient’s clinical needs and currently available therapies.

We have previously shown the feasibility of appendiceal cancer patient-derived tumor organoid models for predicting chemotherapy efficacy.¹¹ Herein, we report using PTOs from appendiceal and colorectal primaries with peritoneal dissemination as a platform to analyze the personalized ex vivo interaction of patient’s specific tumor biology with HIPEC perfusion regimens previously used in prospective randomized trials, including drug type (oxaliplatin vs. MMC), drug dose, and temperature.

PATIENTS AND METHODS

Tumor Biospecimen and Cell Processing

From 23 surgically treated patients with advanced colorectal or appendiceal tumors from June 2019 to January 2020, 23 tissue biospecimens sets were obtained in adherence to previously approved Institutional Review Board (IRB) protocols. Fifteen samples were of appendiceal origin, with nine characterized as low-grade and six as high-grade specimens. The remaining eight were of colorectal origin, with three being high-grade and five low-grade primaries. The specimens were placed in Roswell Park Memorial Institute (RPMI) medium and transferred fresh to the laboratory by a dedicated tissue-procurement manager. Clinical information was not shared with the lab except for the type of tumor and type of prior treatments.

Dissociation and organoid formation were performed within 24 h. Tumors were washed in a solution of Dulbecco’s phosphate-buffered saline (DPBS) with 100 U/mL penicillin/streptomycin, 5 μg/mL gentamicin (G1272; Sigma Aldrich, St. Louis, MS) and 5 μg/mL amphotericin B (A2942, Sigma) for two 5-min washes. Tumors were dissected into pieces measuring less than 2 mm, removing fat and necrotic tissue. Each gram of tissue was placed into a 15-ml conical containing Dulbecco’s modified Eagle’s medium (DMEM) low glucose with no serum, 100,000 cytidine deaminase (CDA) units/mL Collagenase HA 200 Wünch Unit (001–1050; VitaCyte, Indianapolis, IN), 22,000 narcissus pseudonarcissus agglutinin (NPA) units/mL BP Protease (003–1000; VitaCyte), and 20 mM N-acetyl l-cysteine (A9165; Sigma), with the total volume equal to 3 mL solution to 1 g of tumor tissue. The conical was then placed onto a mixing rack and kept at 37 °C until the tumor was dissociated, with a maximum time allowed of 2 h. After the tumor was dissociated, 5 mL of cold RPMI-10 was added to the dissociated tumor to quench the remaining enzymatic activity, and the tumor solution was transferred into a 50-mL conical tube. The tube containing dissociated tumor mixture was then attached to a sterile vacuum filtration kit with 100-μM pore size (SCNY00100; Millipore) to remove remaining undigested pieces of tissue and centrifuged. BD Pharm Lyse™ (555899; BD Biosciences) was added to the cell pellet to perform a red blood cell lyse according to company protocol. Dead cell removal was performed on the remaining cells with a MACS® Dead Cell Removal Kit (130-090-101; Miltenyi Biotec). Remaining cells were counted and ready for use in organoid fabrication.

Organoid Formation

Organoids were formed by encapsulating cells from the dissociated tumor at a cell density of 10 × 10⁶ cells/mL in a thiolated hyaluronic acid (Glycosil; ESI-BIO, Alameda, CA) and methacrylated collagen (PhotoCol®; Advanced BioMatrix, Carlsbad, CA) based hydrogel to a 1:3 ratio. The hydrogel solution/cell suspensions were placed into the wells of a sterile 48-well plate previously coated with cured polydimethylsiloxane (PDMS) in 5 μL (~ 1 mm) volumes. Each deposited hydrogel/cell mixture was then exposed to ultraviolet (UV) light from a BlueWave 75 V.2 UV spot lamp (Dymax Corp., Torrington, CT) for 1 s each. The constructs were then covered with 500 μL media as previously described,¹⁵ with composition of Advanced DMEM-F12 (12634010; ThermoFischer Scientific), 10 mM HEPES buffer solution (15630080; ThermoFisher Scientific), 1 × B-27 supplement (17054044; ThermoFischer Scientific), 1 × N-2 supplement (17502048; ThermoFisher Scientific), 10 mM Nicotinamide (72340; Sigma-Aldrich), 50 ng/mL epidermal growth factor (EGF) (PHG0313; ThermoFisher Scientific), 2 mM Glutamax, 1% penicillin/streptomycin, 1 mM N-acetyl-l-cysteine, 10 μM Y-27632 (S1049), 10 μM SB202190 (S1077; SelleckChem), and 500 nM A83–01 (SML0788; Sigma-Aldrich). Organoids were cultured for 7 days until treatment began.

HIPEC Treatment

After 7 days, media was removed from organoid culture and replaced with 37 °C and 42 °C treatments consisting of control media, 40 mg/3L MMC (M4287; Sigma Aldrich) and 200 mg/m² oxaliplatin (S1224; Selleckchem, Houston, TX) for 120 min in temperature-matched incubators. Similarly, another treatment of 460 mg/m² oxaliplatin was performed for 30 min in matched conditions. PTOs were treated with drug concentrations achieved in the peritoneal cavity of an average patient, with a body surface of 1.9 m², perfused with a volume of 3L of perfusate. After the elapsed time, the treatment solution was removed from each well, and one wash was performed, with fresh media added to replace the wash solution. Organoids were incubated for 72 h before viability assays were performed (Fig. 1).

FIG. 1.

Schematic of tissue and organoid processing and HIPEC regimens performed. a After the establishment of patient organoids and 7 days of culture, chemotherapy at HIPEC concentrations was added at either 37 or 42 °C. Organoids were then washed and allowed to culture for 72 h before viability assays were performed. b Summary of HIPEC regimens performed in this study

Chemotherapy Treatment

Organoids were utilized in chemo-sensitivity screens mimicking intravenous conditions during which these were treated with 5-fluorouracil (5-FU), oxaliplatin, irinotecan, FOLFOX, FOLFIRI, or regorafenib as described previously.¹¹ Stock solutions of 10 mM were prepared for 5-FU (F6627; Sigma), irinotecan (I1406; Sigma), and regorafenib (S1178; Selleckchem) by dissolving in dime-thyl sulfoxide (DMSO), while oxaliplatin and folinic acid (47612; Sigma-Aldrich) were dissolved in deionized H₂O. Treatment solutions were prepared in cell culture media to the following concentrations: 10 μM 5FU, 1 μM oxaliplatin, 1 μM irinotecan, 1:10:1 μM folinic acid:5FU:oxaliplatin (FOLFOX), 1:10:1 μM folinic acid:5FU:oxaliplatin (FOLFIRI), and 1 μM regorafenib. After 7 days, spent media was removed from organoids, and treatments were added to the wells. Organoids were treated for 72 h, after which viability assays were performed.

ATP Viability Assay

After 72 h post-treatment, media was removed from wells containing treated organoids, and 200 μL of 1:1 mixture of media and three-dimensional (3D) CellTiter-Glo® Luminescent Cell Viability Assay solution (G968B; Promega, Madison, WI) was added to each well, with assay performed according to manufacturer’s instructions. The entire contents of each well were added to Costar White Polystrene 96 well Assay Plate (3912; Corning, NY) wells and read on a Variskan™ LUX multimode microplate reader (ThermoFisher Scientific). Values were averaged for experimental groups and analyzed using Graph Pad Prism© v.8 (Graphpad, La Jolla, CA) software.

LIVE/DEAD Analysis

Media was removed from wells containing treated organoids, and viability was assessed by LIVE/DEAD® viability/cytotoxicity kit assays (L3224; Invitrogen, Carlsbad, CA) performed according to manufacturer’s instructions, with organoids incubated for 2 h. Imaging was performed by macroconfocal microscopy (Leica TCS LSI; Leica, Wetzlar, Germany), and composite images were created, with ethidium homodimer-1 red fluorescence indicating dead cell nuclei and calcein AM green fluorescence indicating viable cells.

Statistical Analysis

Data are generally presented as means of number of replicates ± standard deviation. All data are graphed and analyzed for significance using a Student’s t test. Pooled analysis of organoids utilized all individual organoid viability values in matching treatments to compare regimen efficacy. When comparing oxaliplatin regimens, appendiceal sample A1 was not included due to lack of high-oxaliplatin treatment. The authors defined successful treatments as those which reduced tumor viability greater than 50% and reached significance of p < 0.05. For adenosine triphosphate (ATP) assay results, p values were considered significant under 0.05. Data samples were eliminated from the experimental groups if these fell outside two standard deviations from experimental group averages. Sample sizes (n ≥ 5) were determined based on preliminary experiments. These sample sizes, with typically observed standard deviations, allowed statistical significance at p < 0.05 with statistical power greater than 80%.

RESULTS

Organoid Generation

Tissues were successfully dissociated for organoid fabrication, with an average range of 150–200 organoids generated per tumor specimen. This number allowed for the complete setup of the HIPEC screen for comparative analysis of chemotherapy regimens and impact of hyperthermia. Successful HIPEC testing on organoids was performed on 75% (6/8) of colorectal tumor specimens and 73% (11/15) of appendiceal tumor specimens. Two low-grade colon adenocarcinomas as well as two low-grade appendiceal (LGA) and two high-grade appendiceal (HGA) cancers generated a number of PTOs not adequate to support the full HIPEC study size. PTO cells were not expanded to avoid potential genetic drifting and maintain the characteristics of the primary tumor. Three HGA patients (3/4) received systemic chemotherapy prior to cytoreductive surgery (CRS)/HIPEC (FOLFOX n = 2, FOLFIRI/bevacizumab n = 1). Similarly, three LGA patients (3/7) were treated with chemotherapy prior to surgery at the referring facility (FOLFOX/bevacizumab n = 2, FOLFIRI n = 1), while four colon patients (4/6) received systemic chemotherapy prior to CRS/HIPEC (FOLFOX/bevacizumab n = 1, FOLFIRI n = 1, FOLFOX n = 1, FOLFOXIRI/bevacizumab n = 1).

Effect of Hyperthermia on Chemotherapy Efficacy

PTOs were perfused in both 37 °C and 42 °C to quantify the impact of hyperthermia on chemotherapy efficacy. In pooled viability analysis of appendiceal cancer PTOs, perfusion with oxaliplatin at 200 mg/m²/2 h displayed a statistically significant viability reduction when hyperthermia was added (25% vs. 46% p < 0.001). This pooled analysis exhibited by the 200 mg/m² group was not reflective of individual patient responses. Only 5/11 (45%) appendiceal PTO sets exhibited a significant decrease in postperfusion viability with the application of hyperthermia (Table 1). Pooled postperfusion cellular viability analysis of appendiceal cancer PTOs treated with MMC or oxaliplatin at 460 mg/m²/30 min regimen displayed a nonsignificant reduction in viability at 42 °C compared with 37 °C, while only one patient in each group obtained a significant decrease in postperfusion viability by the addition of hyperthermia for these two regimens (Table 1).

TABLE 1.

Treatment response of patient-derived tumor organoids comparing various perfusion protocols. The benefit of adding hyperthermia to the perfusion based on the type of primary and HIPEC regimen. Viability % reflects pooled analysis, while the number of patients (x/n) reflects how many patients exhibited statistically significant benefit from the addition of hyperthermia

Treatment	Appendiceal tumors (n = 11)			Colorectal tumors (n = 6)
	Viability % (x/n)			Viability % (x/n)
	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min
37 °C	30%	46%	63%	57%	49%	75%
42 °C	19% (1/11)	25% (5/11)	54% (1/10)	27% (1/6)	31% (3/6)	53% (1/6)
p value	0.06	< 0.001	0.18	0.02	0.02	0.03

On colorectal PTOs pooled analysis, hyperthermia resulted in a statistically significant decrease in the viability for all three perfusion regimens. Despite this overall decreased pooled viability, only one (17%) colon PTO set exhibited benefit from the addition of heat in both MMC and 460 mg/m² oxaliplatin regimens, while three out of six (50%) patient sets benefited from hyperthermia application in the 200 mg/m² oxaliplatin treatment group (Table 1; Fig. 4). These data suggest that colon cancer PTOs are more sensitive to heat application than appendiceal PTOs. Moreoever, in both appendiceal cancer and colon cancer PTOs, heat potentiates the cytotoxicity of longer-duration low-dose oxaliplatin regimens, while this is not as beneficial for short-duration high-dose oxaliplatin or MMC perfusions.

FIG. 4.

ATP viability assay of a high-grade colon cancer patient (C3) with no significant response to hyperthermia (left panel) along with corresponding LIVE/DEAD imaging (right panel). Green color: live cells; red color: dead cells

HIPEC Treatment Response per Primary

PTOs perfused at 37 °C and 42 °C without the use of chemotherapy did not show statistically significant differences in postperfusion cellular viability (controls, Figs. 2, 3). Pooled data for PTOs perfused with MMC and oxaliplatin 200 mg/m² over 2 h at 42 °C exhibited similar postperfusion cellular viability between the two agents for both appendiceal (19% vs. 25% p = 0.23) and colorectal primaries (27% vs. 31% p = 0.55) (Table 2).

FIG. 2.

Individual patient-derived colorectal tumor organoid results towards HIPEC treatment demonstrating significant variability between patients in terms of response to chemotherapy as well as hyperthermia. C1–C6 colon cancer patients 1–6, LG low grade, HG high grade, CRC colorectal tumor

FIG. 3.

Individual patient-derived appendiceal tumor organoid results towards HIPEC treatment demonstrating significant variability between patients in terms of response to chemotherapy as well as hyperthermia. A1–A11 appendiceal cancer patients 1–11, LG low grade, HG high grade, App appendiceal

TABLE 2.

Treatment response of patient-derived tumor organoids comparing various perfusion protocols per primary. Pooled analysis of PTO viability (%) and individual patient PTO set response (n) based on the type of primary, temperature, and HIPEC regimen. Viability % reflects pooled analysis, while the number of patients (x/n) reflects how many patients of a specific primary exhibited statistically significant cytotoxicity from chemotherapy at 37 or 42 °C

Treatment	Appendiceal tumors (n = 11)			Colorectal tumors (n = 6)
	Viability % (x/n)			Viability % (x/n)
	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min
37 °C	30% (6/11)	45% (5/11)	64% (2/10)	57% (1/6)	49% (0/6)	75% (0/6)
	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p = 0.06
42 °C	19% (10/11)	25% (8/11)	54% (3/10)	27% (4/6)	31% (5/6)	53% (2/6)
	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p < 0.001

However, in appendiceal cancer, when individual patient PTO response was examined, MMC seemed to be more effective than oxaliplatin in decreasing postperfusion tumor viability at both 37 °C [6/11 (55%) vs. 5/11 (45%)] and 42 °C [10/11 (91%) vs. 8/11 (73%)] (Table 2; Fig. 2).

When colorectal PTO response was examined in a patient-by-patient basis, normothermic MMC showed efficacy only in 1/6 (17%) of colorectal PTO sets, while normothermic oxaliplatin at 200 mg/m² did not show efficacy in any patient sets. Colorectal PTO sets from the same patients perfused with hyperthermic MMC and oxaliplatin at 200 mg/m² for 2 h showed a significant reduction in postperfusion viability in 4/6 and 5/6 patient sets, respectively. These data suggest that hyperthermia significantly potentiates both MMC and oxaliplatin cytotoxicity in colorectal tumors (Table 2; Fig. 3).

In colorectal PTO sets, 460 mg/m² oxaliplatin treatment for 30 min was the least effective treatment of all three regimens. Even with the addition of hyperthermia at 42 °C, only two out of six patients (33%) had PTOs exhibiting treatment response (Table 2).

Appendiceal PTOs, HIPEC Response per Grade

LGA PTOs (Table 3) treated with hyperthermic MMC and oxaliplatin at 200 mg/m² demonstrated similar postperfusion cellular viability in both pooled analysis (18% vs. 19%, p = 0.85) and patient-to-patient basis (6/7 vs. 6/7). In HGA primaries, heated MMC resulted in similar postperfusion viability compared with oxaliplatin at 200 mg/m² in pooled PTO analysis (22% vs. 35%, p = 0.11) but displayed better performance in the number of responding PTO sets (4/4 vs. 2/4) (Table 3).

TABLE 3.

Appendiceal cancer PTO response per grade. Pooled analysis of PTO viability (%) and individual patient PTO set response (n) based on the type of primary, temperature, and HIPEC regimen. Viability % reflects pooled analysis, while the number of patients (x/n) reflects how many patients of a specific primary exhibited statistically significant cytotoxicity from chemotherapy at 37 or 42 °C

Treatment	Low grade (n = 7)			High grade (n = 4)
	Viability % (x/n)			Viability % (x/n)
	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min	Mitomycin C 40 mg/2 h	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min
37 °C	30% (5/7)	42% (5/7)	57% (2/7)	32% (1/4)	53% (1/4)	81% (0/3)
	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p = 0.01	p = 0.38
42 °C	18% (6/7)	19% (6/7)	50% (2/7)	22% (4/4)	35% (2/4)	61% (0/3)
	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p = 0.02

Comparison of Oxaliplatin HIPEC Regimens

To investigate how patient tumor organoids can determine the relative efficacy of intraperitoneal chemotherapies as well as the optimal dosing and duration, we analyzed two clinically utilized HIPEC regimens containing oxaliplatin for their cytotoxic effects (Table 4).

TABLE 4.

Treatment response per oxaliplatin regimen based on the type of primary and temperature. Viability (%) reflects pooled analysis, while the number of patients (x/n) reflects how many patients exhibited statistically significant and better response when treated with one drug over the other (superiority analysis) at 37 or 42 °C

Treatment	Appendiceal tumors (n = 10)			Colorectal tumors (n = 6)
	Viability % (x/n)			Viability % (x/n)
	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min	p value	Oxaliplatin 200 mg/m2/2 h	Oxaliplatin 460 mg/m2/30 min	p value
37 °C	46% (5/10)	63% (0/10)	0.01	49% (1/6)	75% (0/6)	< 0.001
42 °C	25% (6/10)	54% (0/10)	0.01	31% (3/6)	53% (0/6)	0.008

In appendiceal cancer PTOs, pooled analysis of normothermic perfusion with 200 mg/m²/2 h oxaliplatin regimen resulted in 46% residual PTO viability, while the viability after 460 mg/m² normothermic oxaliplatin was 63% (p = 0.01). When hyperthermia was applied at 42 °C, the 200 mg/m²/2 h oxaliplatin regimen maintained its cytotoxic superiority over the 460 mg/m²/30 min regimen (25% vs. 54%, respectively, p < 0.001). On the individual patient analysis, the 460 mg/m²/30 min oxaliplatin regimen failed to elicit a meaningful cytotoxic effect in any of the 11 examined appendiceal PTO sets (Table 4).

In colon cancer PTOs, the 460 mg/m²/30 min oxaliplatin regimen was inferior to 200 mg/m²/2 h oxaliplatin under all examined conditions on both pooled postperfusion viability analysis (53% vs. 31%, respectively, p = 0.008) and an individual patient PTO set basis. More specifically, all six colon PTO sets exhibited an inferior residual viability postperfusion with 460 mg/m²/30 min (Table 4).

These data demonstrate that the 460 mg/m²/30 min oxaliplatin regimen is ineffective in both colon and appendiceal PTOs. In addition, this suggests that perfusion with a larger concentration of chemotherapeutic regimens may not be enough to compensate for lack of exposure time (Fig. 4).

Application of Tumor Organoids in Personalized Medicine

Our platform demonstrates the ability to evaluate multiple variables, including treatment time, application of heat, and delivery method. In addition, this allows for the comparison of the effect that a drug has on tumor viability when delivered as systemic chemotherapy and how this effect changes when the same drug is delivered as intraperitoneal chemotherapy with or without heat. As demonstrated in Fig. 5, even though oxaliplatin in concentrations resembling systemic delivery was not effective against PTOs from three appendiceal tumors, the same drug was becoming effective in decreasing viability at concentrations mimicking intraperitoneal delivery.

FIG. 5.

Application of patient-derived tumor organoids (Appendiceal patients A6, A7, A11-central panels) for the comparison of intravenous chemotherapy and HIPEC treatment efficacy. Organoids from patient A6 exhibited resistance to oxaliplatin at systemic concentrations but were sensitive towards oxaliplatin chemotherapy when applied at HIPEC concentrations, as detected by both ATP assays (central panels) and LIVE/DEAD imaging (lateral panels)

DISCUSSION

Significant variation exists in perfusion protocols between HIPEC centers worldwide in regards to type and dosing of chemotherapy as well as achieved range of intraperitoneal temperature.¹⁶ Currently, protocols are designed based on pharmacokinetic and pharmacodynamic data, with area under curve (AUC) ratios utilized as one of the essential determining factors for comparative analyses of various drugs.¹⁶ However, cohort data on kinetics and drug toxicity, by definition, cannot capture the treatment effect of a specific chemotherapy agent on postperfusion residual tumor viability on a patient-by-patient basis.

In the same context, high-dose short-course oxaliplatin (460 mg/m² × 30 min) was studied based on AUC pharmacokinetics and tissue distribution observed in patients undergoing CRS/HIPEC for colon cancer without ex vivo viability data.¹⁷ Subsequently, this protocol was adopted in the design of the Prodige 7 trial that reported inefficacy of oxaliplatin-based CRS/HIPEC (460 mg/m²/30 min) compared with CRS alone to improve overall survival.^18–20 In the current study, the high-dose oxaliplatin regimen exhibited universal viability inferiority and was not successful in achieving 50% cytotoxicity of 1-mm tumor organoids in either appendiceal or colon cancer PTOs. Tumor viability post-2-h perfusion (for both oxaliplatin and mitomycin) was significantly lower compared with 30-min perfusion with high-dose oxaliplatin. In vitro colon cancer cell line studies have also supported the significance of exposure time and showed higher cytotoxicity for oxaliplatin in 2-h compared with 30-min exposure.²¹

The choice of the most effective perfusing agent between oxaliplatin and mitomycin is still a matter of debate. Several retrospective studies had reported comparable outcomes,^2–5 while a study from The American Society of Peritoneal Surface Malignancies suggested superiority of mitomycin in favorable colorectal patients¹ in contrast to another report that demonstrated a survival benefit for oxaliplatin HIPEC in colorectal cancer patients.²² Our group published a randomized trial showing similar oncologic outcomes between the two agents in appendiceal cancer.⁶ Similarly, our PTO model showed no difference in pooled cellular viability for appendiceal cancer and colon cancer between the two HIPEC regimens. Nevertheless, the more important observation is the variations between individual patients suggesting that, for appendiceal cancer, at least at the PTO level, MMC is a more effective drug, while for colon cancer, oxaliplatin performs better in decreasing tumor viability. Our findings suggest that the choice between the two agents should be made at a personalized level based on tumor-derived information.

Hyperthermia is another important factor often administered without unequivocal evidence of efficacy. Initial supporting data emerged from in vitro and animal studies demonstrating a synergistic effect of heat with cytotoxic effects of oxaliplatin;²³ however, other studies failed to establish similar synergism.²⁴ Nonetheless, no comparative analysis has addressed the role of hyperthermia in the clinical setting. Several mechanisms have been proposed to explicate the benefits of hyperthermia with intraperitoneal chemotherapy; Heat-induced impairment of DNA repair, protein denaturation, increased DNA cross-linking efficiency, and apoptosis lead to selective cytotoxicity of tumor cells.^25,26

In contrast, hyperthermia can enhance the expression of heat shock proteins (HSP), which can induce chemoresistance by exerting antiapoptotic and proliferative effects.²⁷ Hyperthermia exploits bidirectional effects on the peritoneal immune environment. While some studies have demonstrated immunotherapeutic effects of local hyperthermia on metastatic disease,²⁸ others reported that thermal immunomodulation is not effective in diminishing the growth of poorly immunogenic tumors.¹⁶ These conflicting findings could explain the heterogeneous impact we observed for hyperthermia. While, in the majority of the patients, hyperthermia potentiated chemotherapy-induced cytotoxicity, in some patients, hyperthermia was not effective in eliciting a similar effect. We also demonstrate that hyperthermia alone is insufficient to elicit tumor cytotoxicity. These observations introduce the possibility that a personalized approach is likely not limited to the personalized use of chemotherapy but also extends to the personalized use of hyperthermia.

As mentioned above, the majority of preclinical studies were conventionally done in cell line cultures in standard dishes [two-dimensional (2D)] conditions and in animal models. These models have several shortcomings to be utilized in clinical settings. Cell cultures in 2D lack tumor microenvironment and do not mimic the organ conformation or structure and have different morphology and physiologic features; thus, these may not be representative of the tumor functionally. Animal models are time-consuming and physiologically diverse compared with the original tumor in human beings. Patient-derived organoids have addressed these concerns substantially and gained significant interest in cancer research in recent years.¹⁷ PTOs are under investigation as platforms for therapeutic prediction and personalized approaches in different cancers.^18–21 To the best of the authors’ knowledge, this is the first study that simultaneously compares different parameters of currently adopted HIPEC regimens in the clinical practice with a personalized approach in a patient-derived 3D model. An important advantage of this approach is testing the tumor cells in the organoids without passage and expansion, therefore, better maintaining their original genetic profiles. The take rates for both tumor types without expansion were 70–80% in a clinically meaningful timeframe. Previous studies have reported substantial similarities in morphology, mutational, and transcriptomic profiles of the original colorectal and liver tumors and their associated PTOs.^18,19 We have previously reported the feasibility of appendiceal cancer PTOs even in low-cellularity tumors and demonstrated that patients’ response to chemotherapy was similar to the response observed in PTOs, supporting the use of organoids as a platform for personalized drug screening.¹¹ We have also demonstrated the feasibility of generating immune-enhanced PTOs for assessment of immunotherapy response by integrating the PTOs with cell populations from lymph nodes.²⁹ The current study further supports the advantages of the PTO platform for a personalized approach to identify optimal treatment options at the patient level over one-size-fits-all conventional approaches. In the future and upon validation, we may be able to see tissue obtained through diagnostic laparoscopy or image-guided biopsy, resulting in organoid development that will determine both systemic and intraperitoneal chemotherapy options.

The modest sample size is one of the limitations to this study. Another valid concern is to demonstrate how well the PTO represents the original tissue. The real question is not absolute structural construct similarity but rather how well the platform would reflect and replicate the innate response in the clinical settings. Specifically, of concern are the maintenance of genetic profiles of the originating tumor tissue. We suspect that our extracellular matrix-derived hydrogel combined with our strategy of maintaining the heterogeneity of the tumor’s cellular microenvironment preserves genomic and phenotypic characteristics of the tumor more effectively than traditional 2D culture on tissue culture on plastic. However, a genomic analysis will be required to test this theory, and these studies are ongoing. These and other comparative analyses are required to address these concerns before preparing PTOs for clinical application.

In conclusion, we show a proof of principle for utilizing patient-derived organoids as a platform to identify the most efficacious intraperitoneal chemotherapy perfusion protocol in a personalized approach. Our model demonstrates that the high-dose short-course oxaliplatin is inferior to the long-course low-dose oxaliplatin as well as mitomycin. Patients have various responses to oxaliplatin versus MMC, and the therapeutic decision should be made at a personalized level to achieve a higher efficacy. While hyperthermia is generally favorable in enhancing cytotoxic effects of HIPEC perfusion, this efficacy is not universal, and some patients might not benefit from hyperthermic perfusion.