FOLFOX is one of the most effective treatments for advanced colorectal cancer. However, cumulative oxaliplatin neurotoxicity often results in halting the therapy. Oxaliplatin functions predominantly via the formation of toxic covalent drug–DNA adducts. We hypothesize that oxaliplatin–DNA adduct levels formed in vivo in peripheral blood mononuclear cells (PBMC) are proportional to tumor shrinkage caused by FOLFOX therapy. We further hypothesize that adducts induced by subtherapeutic “diagnostic microdoses” are proportional to those induced by therapeutic doses and are also predictive of response to FOLFOX therapy. These hypotheses were tested in colorectal cancer cell lines and a pilot clinical study. Four colorectal cancer cell lines were cultured with therapeutically relevant (100 μmol/L) or diagnostic microdose (1 μmol/L) concentrations of [14C]oxaliplatin. The C-14 label enabled quantification of oxaliplatin–DNA adduct level with accelerator mass spectrometry (AMS). Oxaliplatin–DNA adduct formation was correlated with oxaliplatin cytotoxicity for each cell line as measured by the MTT viability assay. Six colorectal cancer patients received by intravenous route a diagnostic microdose containing [14C]oxaliplatin prior to treatment, as well as a second [14C]oxaliplatin dose during FOLFOX chemotherapy, termed a “therapeutic dose.” Oxaliplatin–DNA adduct levels from PBMC correlated significantly to mean tumor volume change of evaluable target lesions (5 of the 6 patients had measurable disease). Oxaliplatin–DNA adduct levels were linearly proportional between microdose and therapeutically relevant concentrations in cell culture experiments and patient samples, as was plasma pharmacokinetics, indicating potential utility of diagnostic microdosing.

This article is featured in Highlights of This Issue, p. 973

Colorectal cancer is the third most frequent cancer and is a leading cause of cancer-related death worldwide (1). Modern treatments for colorectal cancer may include surgery, radiation, chemotherapy, and targeted therapy. Whereas the complete surgical removal of the tumor is considered the mainstay treatment for early tumors (stages I + II), systemic administration of cytotoxic chemotherapy predominates for late-stage disease. Several major advances have been made over the past 15 years resulting in improved overall survival. In large part, these advances are based on the approval of two cytotoxic therapies, oxaliplatin and irinotecan, which can be used in combination regimens with an older standard chemotherapy drug 5-fluorouracil. However, it remains an important need to maximize efficacy and minimize toxicities of those cytotoxic regimens to realize further extension of progression-free survival (PFS) and overall survival (OS) in patients with metastatic colorectal cancer. Since its approval in 2004, FOLFOX, consisting of leucovorin (folic acid), 5-fluorouracil, and oxaliplatin, has been widely adopted as a standard treatment in colorectal cancer in both adjuvant and metastatic settings. Sensory neuropathy is the major dose-limiting toxicity of oxaliplatin and occurs with consecutive and cumulative dose received. Because of function-limiting sensory neuropathy, patients often require dose adjustment or discontinutation of oxaliplatin in the absence of disease progression. Various strategies of intermittent FOLFOX use and FOLFOX rechallenge have been employed to optimize oxaliplatin exposure in metastatic colorectal cancer (1–3). As a result, many patients with colorectal cancer will be exposed and reexposed to oxaliplatin throughout their treatment course until resistance and toxicities develop. A difference in oxaliplatin absorption, metabolism, and DNA repair processes in individual patients potentially alters the accumulation of oxaliplatin in tumor and normal tissue DNA, which may affect efficacy and toxicities of oxaliplatin, respectively. A biomarker to predict the optimal dose of oxaliplatin in an individual patient that maximizes efficacy while potentially minimizing toxicity is a clear unmet need. The cytotoxicity of oxaliplatin is predominantly a consequence of the formation of covalent drug–DNA adducts (Fig. 1; refs. 4–6). Multiple biological pathways such as drug transport, metabolism, DNA damage repair, and cell-cycle modulation affect the activity or efficacy of oxaliplatin (2, 7). Molecular pathway analysis has resulted in a better understanding of these highly complex resistance mechanisms, but these results have yet to be translated into clinical utility.

Figure 1.

Schematically overview of oxaliplatin–DNA adduct formation and methodology used for in vitro or in vivo experiments. A, Simplified scheme for the formation of oxaliplatin-DNA adducts. The 14C label (*) in oxaliplatin is located in the DACH carrier group, therefore both forms of adducts (mono- and diadducts) can be detected by AMS. When adducts are repaired, the 14C label is removed from the DNA and no longer be detected. B and C, Illustration of experimental design of in vitro or in vivo dosing and sensitivity correlation. B, Colorectal cancer cell cultures were exposed to [14C]oxaliplatin and drug–DNA adduct level were assessed via AMS. Drug sensitivity was determined via MTT and IC50 values were correlated to adduct level. C, Patients with colorectal cancer were administered a [14C]oxaliplatin microdose prior to blood sampling and FOLFOX chemotherapy. DNA was then purified from isolated PBMC and assayed for drug–DNA adduct formation via AMS. The drug–DNA adduct levels were correlated to tumor reduction.

Figure 1.

Schematically overview of oxaliplatin–DNA adduct formation and methodology used for in vitro or in vivo experiments. A, Simplified scheme for the formation of oxaliplatin-DNA adducts. The 14C label (*) in oxaliplatin is located in the DACH carrier group, therefore both forms of adducts (mono- and diadducts) can be detected by AMS. When adducts are repaired, the 14C label is removed from the DNA and no longer be detected. B and C, Illustration of experimental design of in vitro or in vivo dosing and sensitivity correlation. B, Colorectal cancer cell cultures were exposed to [14C]oxaliplatin and drug–DNA adduct level were assessed via AMS. Drug sensitivity was determined via MTT and IC50 values were correlated to adduct level. C, Patients with colorectal cancer were administered a [14C]oxaliplatin microdose prior to blood sampling and FOLFOX chemotherapy. DNA was then purified from isolated PBMC and assayed for drug–DNA adduct formation via AMS. The drug–DNA adduct levels were correlated to tumor reduction.

Close modal

This complexity motivated us to investigate the possible utilization of microdose-induced drug–DNA adducts as functional biomarkers for oxaliplatin response in colorectal cancer. The subtherapeutic “diagnostic microdosing” approach has been reported to potentially provide predictive information regarding chemosensitivity to platinum-based chemotherapy (5, 8, 9). This approach uses the tracing of nontoxic diagnostic microdoses of 14C-labeled oxaliplatin via accelerator mass spectrometry (AMS), which has attomole (10−18 mole) sensitivity for 14C; easily sensitive enough for clinical applications. AMS provides an isotope ratio for each DNA sample that allows calculation of the drug concentration in tissues, blood, protein, or nucleic acids at concentrations that are difficult or impossible to measure with other techniques. AMS therefore enables human studies with radiolabeled drug while circumventing the need for toxic drug or high radiation exposures (10, 11). We hypothesize that DNA damage caused by a single subtoxic microdose of oxaliplatin can predict patient outcomes such as tumor shrinkage and survival. We report herein investigation of oxaliplatin microdosing with four colorectal cancer cell lines and 6 patients with colorectal cancer. The goal of the project was to define the appropriate chemical and radiochemical dose for the microdose composition, to establish protocols for the procedure, and to gather preliminary clinical data in support for a larger clinical trial.

Chemicals

Unlabeled oxaliplatin (5 mg/mL) was purchased from Sanofi-Aventis. 14C-labeled oxaliplatin [specific activity 54 mCi/mmol with the 14C-label in the diaminocyclohexane (DACH) group] was obtained from Moravek Biochemicals. [14C]oxaliplatin for injection was prepared under good manufacturing practice (GMP) at the GMP facility at UC Davis (Sacramento, CA). The [14C]oxaliplatin drug substance was dissolved with sterile water for injection (APP Pharmaceuticals). The solution was filter sterilized with a 0.2-μm PES syringe filter and the resulting drug product was stored at -20°C. Specific activity was determined by liquid scintillation counting (LSC). Mixtures of 14C-labeled and unlabeled drug were used to minimize the usage of radiocarbon and achieve the different specific activities required for microdoses and therapeutic doses. Drug solutions for the indicated experiments were prepared immediately before use.

Cell lines and cytotoxicity assay

Four human colorectal cancer cell lines (CLL-228, CLL-229, CRL-2134, and HBT-38) were purchased from the ATCC and cultured with the recommended medium without additional verification of identity. The IC50 values were determined in triplicate via standard MTT assay as described elsewhere (12). The MTT assay and diagnostic microdoses were performed on cell lines with as close as possible to the same number of passages. Therefore, cell cultures were continuously exposed to oxaliplatin for 72 hours before assay development.

Drug treatment and AMS analysis

Using a protocol derived from previously reported [14C]carboplatin microdosing studies (9, 8, 13), cells were cultured to >90% confluence, dosed with indicated doses of [14C]oxaliplatin, further cultured for up to 24 hours, and subjected to DNA isolation and AMS analysis as described previously (13). Briefly, cells were seeded at 1 × 106 cells per 60-mm dish and allowed to attach overnight at 37°C and 5% CO2 in a humidified atmosphere. At hour 0, cells were dosed with 1 μmol/L (microdose) or 100 μmol/L oxaliplatin (therapeutic dose), each supplemented with [14C]oxaliplatin at 500 dpm/mL. Cultures were incubated for 4 hours and then washed twice with PBS and maintained thereafter with drug-free culture media for indicated periods of time to mimic the human oxaliplatin plasma half-life (0.5–16 hours) and capture possible DNA repair kinetics (14). Cells were harvested at 0, 2, 4, 8, and 24 hours after initiation of dosing. DNA was extracted from collected cells with the Wizard Genomic DNA Purification Kit according to the manufacturer's instructions (Promega). DNA quantity and quality was determined with a NanoDrop 1000 spectrophotometer, the purity was ensured by obtaining a 260/280 nm OD ratio of approximately 1.9. The protocol included washing of the DNA pellet with aqueous alcohol solutions to eliminate contamination by proteins and small-molecule metabolites. Assessing the 260/280 nm OD ratio minimized the possibility of contamination of the samples with RNA. All DNA samples were submitted to Lawrence Livermore National Laboratory (LLNL) for AMS analysis of radiocarbon content using an established protocol (15). Ten micrograms of DNA per sample was converted to graphite and measured by AMS for 14C quantification as described previously. Triplicate sets of AMS experiments were performed for each cell line and time point. The data was plotted as oxaliplatin-DNA adducts per 108 nucleotides (nt) over time.

A pilot feasibility diagnostics clinical trial

The study titled “Pilot Study of a Carbon 14 Oxaliplatin Microdosing Assay to Predict Exposure and Sensitivity to Oxaliplatin-Based Chemotherapy in Advanced Colorectal Cancer” (ClinicalTrials.gov identifier NCT02569723) was a feasibility study of the diagnostic microdosing approach and was approved by the UC Davis Institutional Review Board and conducted under an IND from FDA. The study was conducted in accordance with the Declaration of Helsinki and performed after UC Davis Institutional Review Board approval and after obtaining written informed consent from the subjects. The study included patients with locally advanced or metastatic colorectal adenocarcinoma. The study design was derived from a previous carboplatin-based microdosing pilot clinical study that accrued patients with bladder and lung cancer, but with an emphasis on correlation of microdose-induced drug–DNA adduct levels to mean tumor volume change rather than response as defined by RECIST criteria (9, 8). The therapeutic dose of oxaliplatin was administered at 85 mg/m2. Diagnostic microdoses of oxaliplatin were administered to patients at 1% of the therapeutic dose. [14C]oxaliplatin was given at 2 × 106 dpm/kg. The 14C containing microdose and therapeutic dose were administered by a 2-hour intravenous infusion. Toxicity of the two administered dose levels (microdose and therapeutic dose) was assessed using Common Terminology Criteria for Adverse Events (CTCAE). Patient response to FOLFOX chemotherapy was evaluated using the RECIST and correlated to oxaliplatin–DNA adduct levels. Unlabeled oxaliplatin and [14C]oxaliplatin were administered unmixed but simultaneously, and peripheral blood specimens from a separate access point were drawn into BD Vacutainer CPT tubes with sodium citrate (Becton Dickinson). Peripheral blood mononuclear cells (PBMC) were isolated within 2 hours of collection by centrifugation according to the manufacturer's instructions. A proportion of ultra-centrifuged (10 kDa MWCO) and total plasma were used for liquid scintillation counting (LSC).

LC/MS

Human plasma samples were made for establishing a calibration curve. For each sample, 50 μL of oxaliplatin (Selleckchem.com) solution was prepared at concentrations of 0.005, 0.01, 0.05, 0.1, 0.5, 1, 5, 10, and 50 μg/mL in H2O (Thermo Fisher Scientific), and were each spiked with 450 μL of pooled blank human plasma in sodium citrate buffer (Bioreclamation IVT). Quality control (QC) solutions were made by mixing of 50 μL of oxaliplatin stock solution at 0.025, 0.25, and 25 μg/mL in H2O with 450 μL of blank human plasma. The resulting human plasma calibration standards, quality control samples, and the patient plasma samples were filtered with 10 kDa MWCO ultrafiltration filters (EMD Millipore) prior to LC/MS analysis. Twenty microliters of each resulting filtrate was spiked with 80 μL of 1 μg/mL carboplatin (Hospira Inc) in H2O, which served as an internal standard. Five microliters of the resulting mixture was injected into a Waters Acquity UPLC with a BEH C18 1.7 μm, 2.1 mm × 50 mm column. The following mobile phases were used: (A) H2O with 0.1% formic acid (Thermo Fisher Scientific), and (B) methanol (Thermo Fisher Scientific) with 0.1% formic acid. Mobile phase A was also used for purging and needle washing between each injection. The flow rate was 0.4 mL/minute and the column temperature was 50°C with an autosampler temperature of 10°C. The output of the UPLC was fed to a Waters Xevo TQ-S triple quadrupole MS/MS system, which was used to ionize target molecules and monitor the ion m/z fragmentation transitions from 398.3 → 96.1 for oxaliplatin quantification, and 371.9 → 294.2 for carboplatin at multiple reaction monitoring (MRM) mode. The retention times were 1.02 and 0.70 minutes for oxaliplatin and carboplatin, respectively. The detection range for oxaliplatin was from 0.5 to 5,000 ng/mL. The calibration curve was modeled with a weighted (1/x2) least-squares linear regression algorithm. The lower limit of quantification (LLOQ) was 0.5 ng/mL. The extraction yield for oxaliplatin was 73.67% ± 8.12% and the matrix effect enhanced the oxaliplatin MS signal by 16.71%. Both inter- and intra-batch accuracy were lower than 10% (% deviation) and both intra- and inter-batch precision were also lower than 10% (%CV).

Statistical analysis

Statistical analysis was performed using GraphPad Prism software (GraphPad Software Inc.) using a two-tailed t test or a one-way ANOVA with Bonferroni post hoc test, as appropriate. A P value below 0.05 was considered statistically significant. All experiments were carried out at least in triplicate to enable statistically significant comparisons of the results. All results are expressed as the mean ± SD unless otherwise noted. A simple correlation of the adduct levels and tumor volume change is reported.

Oxaliplatin–DNA adduct levels induced by diagnostic microdose and therapeutic concentrations in colorectal cancer cell cultures are dose proportional

We tested the hypothesis that the levels of microdose-induced oxaliplatin–DNA adducts in colorectal cancer cell cultures are proportional to those induced by therapeutic doses. Four colorectal cancer cell lines were exposed to therapeutically relevant (100 μmol/L) or diagnostic microdose (1 μmol/L) concentrations of [14C]oxaliplatin for 4 hours, followed by incubation in drug-free medium. Cells were harvested at various time points over a period of 24 hours (Figs. 1 and 2) and drug–DNA adduct levels were determined through AMS analysis of purified genomic DNA. The dosing protocol was designed to crudely mimic in vivo cell exposure to oxaliplatin (exponential decrease of drug concentration in blood over approximately a day). As expected, there was a time-dependent increase in oxaliplatin–DNA adduct levels during the first 4 hours (Fig. 2) of drug exposure to either concentration followed by a gradual decrease over the subsequent 20 hours of cell incubation in drug-free medium. The peak-adduct levels varied considerably between the different cell lines even though all cell lines showed the same overall trend of adduct concentration peak and decline over time. The mean oxaliplatin–DNA adduct levels of across all cell lines from microdose administration (1–10 adducts/108 nt) were approximately 100-fold lower than the levels obtained from therapeutic dosing (100–1,000 adducts/108 nt, Fig. 2C). Linear regression analysis showed that the adduct levels induced by the two oxaliplatin dose concentrations were highly linear and correlated significantly (R2 = 0.95, P < 0.0001, Fig. 2D), which supports the concept that microdosing can be used to predict the DNA adduct levels induced by therapeutic oxaliplatin without negatively impacting cell viability.

Figure 2.

Oxaliplatin–DNA adducts induced by microdose and therapeutic concentrations in colorectal cancer cell cultures are dose proportional. Indicated four colorectal cancer cells lines were dosed for 4 hours followed 20-hour incubation in drug-free medium. Microdoses (1 μmol/L; A) or therapeutic dose (100 μmol/L; B) induced oxaliplatin–DNA adduct level over time. C, Dose proportionality of microdose and therapeutic dose induced oxaliplatin–DNA adduct level (log scale). D, Linear correlation of oxaliplatin–DNA adduct level induced by microdosing and therapeutic oxaliplatin, showing that the adduct levels induced were highly linear (R2 = 0.91, P < 0.0001). Mean and SE are shown.

Figure 2.

Oxaliplatin–DNA adducts induced by microdose and therapeutic concentrations in colorectal cancer cell cultures are dose proportional. Indicated four colorectal cancer cells lines were dosed for 4 hours followed 20-hour incubation in drug-free medium. Microdoses (1 μmol/L; A) or therapeutic dose (100 μmol/L; B) induced oxaliplatin–DNA adduct level over time. C, Dose proportionality of microdose and therapeutic dose induced oxaliplatin–DNA adduct level (log scale). D, Linear correlation of oxaliplatin–DNA adduct level induced by microdosing and therapeutic oxaliplatin, showing that the adduct levels induced were highly linear (R2 = 0.91, P < 0.0001). Mean and SE are shown.

Close modal

Genomic drug–DNA adduct levels correlate with colorectal cancer cell sensitivity to oxaliplatin

We next evaluated whether the sensitivity of colorectal cancer cell cultures to oxaliplatin can be determined by analysis of drug–DNA adduct formation. Cell sensitivity toward oxaliplatin was determined via the MTT viability assay after a 72-hour oxaliplatin exposure. The four tested cell lines were allocated into either sensitive or the resistant groups (two cell lines in each group based on relative oxaliplatin sensitivity). At each time point, the two most resistant cell lines (CLL-228 and CLL-229) had lower oxaliplatin–DNA adduct levels than the two most sensitive cell lines (CRL-2134 and HTB-38; Fig. 3A and B; Table 1). There was a significant difference between the two groups after 4-hour exposure to the microdose (P < 0.01) or therapeutic dose (P < 0.05). The mean area under the adduct curve (AUCadducts) of the two resistant cell lines was lower than the AUCadducts of the sensitive cell lines but did not reach significance (Supplementary Fig. S1).

Figure 3.

Correlation between colorectal cancer cell line oxaliplatin sensitivity and oxaliplatin–DNA adduct levels. Box plot (whiskers, min–max) comparing adduct levels in four colorectal cancer cell lines over 24 hours. Compared with the two most resistant cell lines CLL-228 and CLL-229 (gray box), the two most sensitive cell lines CRL-2134 and HTB-38 (white box) had significant (**, P < 0.01; *, P < 0.05, one-way ANOVA with Bonferroni multiple comparison test) higher adduct level after 4 hours of microdose (A) or therapeutic dose oxaliplatin exposure (B).

Figure 3.

Correlation between colorectal cancer cell line oxaliplatin sensitivity and oxaliplatin–DNA adduct levels. Box plot (whiskers, min–max) comparing adduct levels in four colorectal cancer cell lines over 24 hours. Compared with the two most resistant cell lines CLL-228 and CLL-229 (gray box), the two most sensitive cell lines CRL-2134 and HTB-38 (white box) had significant (**, P < 0.01; *, P < 0.05, one-way ANOVA with Bonferroni multiple comparison test) higher adduct level after 4 hours of microdose (A) or therapeutic dose oxaliplatin exposure (B).

Close modal
Table 1.

Relevant cell line data for oxaliplatin IC50, microdose (micro), and therapeutic dose (thera) induced oxaliplatin–DNA adduct level after 4 and 24 hours, area under the adduct curves (AUC adduct) induced by repair rates in four colorectal cancer cell lines.

DNA Damage:oxaliplatin adducts/108 ntAUC Adducts/108 nt per hour8–24–Hour repair adducts/108 nt per hour
OxaliplatinTheraMicroTheraMicroTheraMicro
Cell lineIC50 (μmol/L)4 Hours24 Hours4 Hours24 Hours
CLL-228 102 63,134 3,051 66.10 25.67 111,114 945.1 197 1.39 
CLL-229 17.8 14,651 8,159 129.6 68.82 253,169 2,180 302 2.66 
CRL-2134 0.9 9,640 5,239 96.66 36.54 161,558 1,414 173 2.13 
HTB-38 0.3 69,952 47,821 851.8 312.1 1,220,000 9,769 523 6.53 
DNA Damage:oxaliplatin adducts/108 ntAUC Adducts/108 nt per hour8–24–Hour repair adducts/108 nt per hour
OxaliplatinTheraMicroTheraMicroTheraMicro
Cell lineIC50 (μmol/L)4 Hours24 Hours4 Hours24 Hours
CLL-228 102 63,134 3,051 66.10 25.67 111,114 945.1 197 1.39 
CLL-229 17.8 14,651 8,159 129.6 68.82 253,169 2,180 302 2.66 
CRL-2134 0.9 9,640 5,239 96.66 36.54 161,558 1,414 173 2.13 
HTB-38 0.3 69,952 47,821 851.8 312.1 1,220,000 9,769 523 6.53 

A pilot diagnostic feasibility trial in human patients with colorectal cancer

Our main objective was to test the feasibility of using [14C]oxaliplatin diagnostic microdoses to enable prediction of therapeutic oxaliplatin exposure, pharmacodynamics (drug–DNA adduct levels), and corresponding tumor volume changes. The study utilized the assessment of oxaliplatin–DNA adduct levels as biomarkers in genomic DNA from PBMCs as surrogates for tumor biopsy samples and also examined the intrapatient variability of micro- and therapeutic dose pharmacokinetics. We tested the hypotheses that the intravenous infusion of a diagnostic [14C]oxaliplatin microdose will predict the pharmacokinetic exposure to treatment dose oxaliplatin and that accumulation of the microdose-induced oxaliplatin–DNA damage levels in PBMCs will correlate with patient response to subsequent full-dose oxaliplatin-based chemotherapy.

The clinical feasibility trial was initiated at UC Davis Comprehensive Cancer Center (ClinicalTrials.gov identifier NCT02569723) and a total of 6 patients with locally advanced or metastatic colorectal adenocarcinoma were accrued. The patients that completed the diagnostic microdosing and blood sampling, also received an additional [14C]oxaliplatin dose with their subsequent full-dose FOLFOX, which is referred to herein as a therapeutic dose of [14C]oxaliplatin.

Similar to what was observed clinically in bladder and non–small cell lung cancer (9, 8), the microdose composition of [14C]oxaliplatin of 2 × 106 dpm/kg and a total oxaliplatin dose of 1% of the therapeutic dose, the 14C-signal in DNA isolated from PBMCs was approximately 10–100 times the background, which allowed accurate adduct measurement by AMS. As expected, no microdose-associated adverse events were observed in any of the patients. The administered diagnostic [14C]oxaliplatin microdose appears to be safe in this patient population. In comparison to the annual effective radiation dose equivalent from natural internal sources of 1.6 mSv per person (16), and a radiation exposure for an abdominal CT scan of 10 mSv (17), the average administration of the diagnostic microdose was not greater than 0.01 mSv.

Linear correlation of oxaliplatin pharmacokinetics between microdose and therapeutic dose

Immediately following the blood draws, whole plasma was isolated for pharmacokinetic analysis via liquid scintillation counting (LSC) and LC/MS. The total and free oxaliplatin plasma concentration after diagnostic microdoses (Fig. 4AC) and therapeutic doses (Fig. 4DF) showed the expected increase over the 2-hour intravenous infusion and an approximately biphasic elimination kinetic in the following 22 to 46 hours. At 24 hours after dosing, only 50%–75% of total 14C label was cleared from whole plasma (Fig. 4A and D). Analysis of free [14C]oxaliplatin after removal of plasma proteins by 10 kDa MWCO filtration showed that at 24 hours less than 10% of the [14C]oxaliplatin was detectable (Fig. 4B and E). The expected discrepancy between the 14C concentrations in total and ultra-centrifuged plasma has been previously reported and is closely linked to oxaliplatin's high binding probability to plasma proteins and erythrocytes (14). To further validate the approach of determining [14C]oxaliplatin PK with LSC, we used LC/MS to determine the oxaliplatin concentration in 10 kDa MWCO filtrated patient plasma specimen (Fig. 4C and G). Analysis of free oxaliplatin plasma concentrations after micro- or therapeutic doses via LSC (Fig. 4D) and LC/MS (Fig. 4H) shows a highly significant linear correlation (R2 = 0.94, P < 0.0001 and R2 = 0.68, P < 0.0001, respectively). As expected, the LC/MS method was less sensitive compared with LSC and AMS, because several time points were below the lower limit of detection due to the very low (subtherapeutically) oxaliplatin concentration in the diagnostic microdose.

Figure 4.

Oxaliplatin plasma concentration after microdose and therapeutic dose infusion in patients with colorectal cancer are dose proportional. Indicated 3 patients with colorectal cancer (patient #1: Pt1 , #2: Pt2 , and #3: Pt3 ) were infused for 2 hours with a microdose (A–C) or therapeutic dose of oxaliplatin (E–G). Whole plasma oxaliplatin concentrations were analyzed over a period of 24 hours (A) or 48 hours (E) via LSC. Free oxaliplatin concentrations from 10 kDa MWCO centrifuged plasma was determined via LSC (B and E) or LC/MS (C and F). Mean values and SD are shown. Linear correlation of free plasma oxaliplatin after microdose and therapeutic dose determined via LSC (D) or LC/MS (H).

Figure 4.

Oxaliplatin plasma concentration after microdose and therapeutic dose infusion in patients with colorectal cancer are dose proportional. Indicated 3 patients with colorectal cancer (patient #1: Pt1 , #2: Pt2 , and #3: Pt3 ) were infused for 2 hours with a microdose (A–C) or therapeutic dose of oxaliplatin (E–G). Whole plasma oxaliplatin concentrations were analyzed over a period of 24 hours (A) or 48 hours (E) via LSC. Free oxaliplatin concentrations from 10 kDa MWCO centrifuged plasma was determined via LSC (B and E) or LC/MS (C and F). Mean values and SD are shown. Linear correlation of free plasma oxaliplatin after microdose and therapeutic dose determined via LSC (D) or LC/MS (H).

Close modal

Correlation of microdose-induced oxaliplatin–DNA monoadducts in PBMC with tumor response to subsequent full-dose FOLFOX chemotherapy

A key aspect of this study is that the level of oxaliplatin–DNA adducts in PBMCs was used as a surrogate biomarker for analysis of tumor DNA. This approach was chosen due to limited access to biopsy samples from patients with metastatic colorectal cancer. This choice is also justified by several studies by us and others that showed that platinum-based drug–DNA adducts in PBMCs correlate with tumor response to chemotherapy (5, 6, 8, 9, 18–22). Even though most of the free [14C]oxaliplatin was cleared from blood within 24 hours, adduct levels were readily measurable for up to 48 hours. The oxaliplatin–DNA adduct levels following the diagnostic microdose or therapeutic dose ranged from 0.1–6 to 13–1,047 adducts per 108 nucleotides, respectively (Fig. 5A and B; Table 2). Mean oxaliplatin–DNA adduct levels across all microdosed patients were approximately 100-fold lower than those observed after therapeutic dosing (Fig. 5C). Linear regression analysis showed that the adduct levels induced by the two oxaliplatin dose concentrations were highly linear and correlated significantly (R2 = 0.67, P < 0.0001, Fig. 5D), suggesting that the nontoxic diagnostic microdosing approach can be used to predict the DNA adduct levels induced by therapeutic oxaliplatin without negatively impacting patient health or PBMC viability.

Figure 5.

Microdose and therapeutic dose–induced oxaliplatin–DNA adduct level are dose proportional and correlate with FOLFOX-induced tumor reduction. Indicated three patients with colorectal cancer (patient #1: Pt1 , #2: Pt2 , and #3: Pt3 ) were infused for 2 hours with a microdose (A) or therapeutic dose of oxaliplatin (B). Oxaliplatin–DNA adduct levels were analyzed over a period of 24 or 48 hours via AMS. C, Dose proportionality of log microdose and therapeutic dose–induced oxaliplatin–DNA adduct level (mean values and SE are shown). D, Linear correlation of oxaliplatin–DNA adduct level induced by a diagnostic micro- or a therapeutic dose of oxaliplatin (R2 = 0.63, P < 0.0001). Correlation of microdose (R2 = 0.82, P = 0.033; E) or therapeutic dose (R2 = 0.65, P = 0.099; F) induced oxaliplatin–DNA adduct level with mean tumor volume change after patient received 3–4 cycles of FOLFOX chemotherapy.

Figure 5.

Microdose and therapeutic dose–induced oxaliplatin–DNA adduct level are dose proportional and correlate with FOLFOX-induced tumor reduction. Indicated three patients with colorectal cancer (patient #1: Pt1 , #2: Pt2 , and #3: Pt3 ) were infused for 2 hours with a microdose (A) or therapeutic dose of oxaliplatin (B). Oxaliplatin–DNA adduct levels were analyzed over a period of 24 or 48 hours via AMS. C, Dose proportionality of log microdose and therapeutic dose–induced oxaliplatin–DNA adduct level (mean values and SE are shown). D, Linear correlation of oxaliplatin–DNA adduct level induced by a diagnostic micro- or a therapeutic dose of oxaliplatin (R2 = 0.63, P < 0.0001). Correlation of microdose (R2 = 0.82, P = 0.033; E) or therapeutic dose (R2 = 0.65, P = 0.099; F) induced oxaliplatin–DNA adduct level with mean tumor volume change after patient received 3–4 cycles of FOLFOX chemotherapy.

Close modal
Table 2.

Pharmacokinetic, pharmacodynamic, and response data for 3 patients with colorectal cancer given [14C]oxaliplatin (Ox) as a microdose (Micro) or therapeutic dose (Thera) as part of the FOLFOX chemotherapy regimen.

2-Hour peakAUC serumOxaliplatin–DNA level: adducts/108 ntAUC: Adducts/Repair: adducts/108 nt per hourChemotherapy
serum oxaliplatin:oxaliplatin:In vivoIn vitro108 ntIn vivoIn vitroresponse after
Pt#Doseng/mL (Cmax)ng/mL per hour4 Hours24 Hours24 Hoursper hour3–4 cyclesa
Micro 12.4 68.47 3.068 1.242 — 48.83 −0.091 — PR (−28%) 
 Thera 1,175 9,732 217.8 345.8 — 11,453 +6.396 —  
Micro 15.7 80.02 5.963 2.824 — 102.0 −0.157 — PR (−72%) 
 Thera 1,736 10,519 1,047 617.8 — 31,228 −21.48 —  
Micro 11.5 86.36 3.727 2.460 4.888 72.58 −0.063 +0.059 PR (−41%) 
 Thera 1,792 18,215 548.2 649.1 465.4 24,221 +5.045 −4.143  
Micro 13.4 108.9 2.140 1.400 1.472 42.34 −0.037 −0.033 Not evaluable 
 Thera 1,142 11,313 287.1 238.8 258.0 9,872 −2.412 −1.453  
Micro 8.6 66.06 2.719 1.536 1.938 50.73 −0.059 −0.039 PR (−41%) 
 Thera 973.9 7,550 172.8 286.3 121.9 9,568 +5.475 −2.544  
Micro 14.3 114.8 2.589 2.095 1.071 56.51 −0.025 −0.076 PD (+9%) 
 Thera 1,254 15,725 315.2 200.0 305.7 10.513 −5.760 −0.476  
2-Hour peakAUC serumOxaliplatin–DNA level: adducts/108 ntAUC: Adducts/Repair: adducts/108 nt per hourChemotherapy
serum oxaliplatin:oxaliplatin:In vivoIn vitro108 ntIn vivoIn vitroresponse after
Pt#Doseng/mL (Cmax)ng/mL per hour4 Hours24 Hours24 Hoursper hour3–4 cyclesa
Micro 12.4 68.47 3.068 1.242 — 48.83 −0.091 — PR (−28%) 
 Thera 1,175 9,732 217.8 345.8 — 11,453 +6.396 —  
Micro 15.7 80.02 5.963 2.824 — 102.0 −0.157 — PR (−72%) 
 Thera 1,736 10,519 1,047 617.8 — 31,228 −21.48 —  
Micro 11.5 86.36 3.727 2.460 4.888 72.58 −0.063 +0.059 PR (−41%) 
 Thera 1,792 18,215 548.2 649.1 465.4 24,221 +5.045 −4.143  
Micro 13.4 108.9 2.140 1.400 1.472 42.34 −0.037 −0.033 Not evaluable 
 Thera 1,142 11,313 287.1 238.8 258.0 9,872 −2.412 −1.453  
Micro 8.6 66.06 2.719 1.536 1.938 50.73 −0.059 −0.039 PR (−41%) 
 Thera 973.9 7,550 172.8 286.3 121.9 9,568 +5.475 −2.544  
Micro 14.3 114.8 2.589 2.095 1.071 56.51 −0.025 −0.076 PD (+9%) 
 Thera 1,254 15,725 315.2 200.0 305.7 10.513 −5.760 −0.476  

Note: Response was determined according to RECIST criteria as either partial response (PR) or one almost complete response (CR) or progressive disease (PD).

Abbreviation: Pt#, patient number.

aNumbers represent percent tumor shrinkage after last cycle.

The clinical response of the patients with colorectal cancer was determined by measuring the percent change in the sum of the target lesions (measurable disease) and is shown in Table 2. Measurable disease data were collected from 5 of the 6 patients accrued. Correlation analysis of the mean change in target lesions after 3-4 cycles of FOLFOX with induced microdose (Fig. 5E) or therapeutic dose (Fig. 5F) adduct levels at 4 hours shows a linear relationship (R2 = 0.082, P = 0.033, R2 = 0.62, P = 0.099, respectively). Patients responded to FOLFOX chemotherapy to varying degrees with the patient achieving the highest 4-hour adduct-level yielding almost complete remission of the measured target lesions after 12 cycles of therapy. In contrast, the patient with the lowest 4-hour adduct level had initially disease progression after 3 cycles as evidenced by a 9% increase in mean tumor volume but showed a modest tumor volume reduction of 6% after 8 cycles. In addition to yielding the best correlations, 4-hour time point post microdose is reasonable for clinical implementation of such a diagnostic test because patients would not have to make repeat hospital visits in between dosing and blood sampling. This early trend associating oxaliplatin–DNA adduct level with colorectal cancer patient response is encouraging. However, more patients need to be accrued and the protocol needs to be optimized in order to maximize accrual and more closely match how such a test would be utilized in the clinic.

The results presented herein indicate that (i) oxaliplatin–DNA adduct levels in colorectal cancer cell lines correlate to cellular sensitivity to the drug, (ii) that microdose-induced oxaliplatin–DNA adduct levels in PBMC are predictive of those formed by therapeutic doses in vivo, and (iii) that such data are predictive of FOLFOX response, at least for some time points. Collectively, these data point to potential clinical utility for diagnostic microdosing.

Patients with colorectal cancer often respond well to FOLFOX only to discontinue treatment due to cumulative and irreversible oxaliplatin-induced neuropathy. Diagnostic microdosing has the potential to enabled adjustment of the oxaliplatin dose or the frequency of each cycle to target a adduct levels at specific time points or the area under the adduct curve that would enable greater precision FOLFOX. For example, patients that form relatively high oxaliplatin–DNA adducts could be given smaller doses or less frequent cycles of FOLFOX to avoid early-onset neuropathy while maintaining efficacy. For those patients with relatively low oxaliplatin–DNA adducts, dose-dense therapy or alternative chemotherapeutic regimens could be considered to enhance efficacy.

Tumor-related oxaliplatin resistance is based on a multitude of molecular mechanisms, including alterations in drug transport, detoxification, DNA damage response and repair, cell death (apoptotic and nonapoptotic), and epigenetic mechanisms. Diagnostic microdosing is an attractive concept due to the simplicity of measuring a single phenotypic marker that is influenced by a large array resistance factors, and without the need for a tumor biopsy. Furthermore, measurement of drug–DNA adduct levels may be useful as a tool for better understanding DNA repair pathways of clinical relevance.

We determined the feasibility of a diagnostic microdosing approach for the study of colorectal cancer cellular sensitivity to oxaliplatin. Our data support the concept that microdose-induced oxaliplatin–DNA adduct levels are predictive of cellular sensitivity to oxaliplatin and tumor volume control in patients with colorectal cancer. This approach is an extension our previous work that focused on bladder and lung cancer, predominantly for cisplatin- and carboplatin-based therapy response prediction, but with extension of the methodology for dosing, sample collection, and data analysis for assessment of FOLFOX optimization (9, 8, 13, 23). There was a significant linear correlation between oxaliplatin–DNA adduct levels induced by microdose and therapeutic dose in four colorectal cancer cell lines. In addition, cell line sensitivity to oxaliplatin significantly correlated to oxaliplatin–DNA adduct levels after 4 hours of drug exposure with both drug concentrations.

Our preclinical cell line data formed the foundation of the feasibility microdosing trial aimed at determining whether oxaliplatin–DNA adduct levels in PBMCs correlate with the response to FOLFOX chemotherapy in patients with colorectal cancer. Our pilot study showed that a diagnostics feasibility study can be performed in patients without any detectable toxicity associated with the microdose. This pilot study was not designed to demonstrate statistical significance of drug–DNA adduct frequency as a biomarker, but still yielded some encouraging results (Fig. 5).

First, we developed protocols for conducting a clinical trial with a 14C-labeled drug and processing samples for AMS analysis. On the basis of this effort, we decided to perform liquid biopsies (blood sampling) for up to 48 hours after dosing. At 24 hours after dosing, there were measurable yet large interpatient variations in drug–DNA adduct levels, while the [14C]oxaliplatin concentration in plasma was negligible by LC/MS.

Second, we defined a clinically useful of oxaliplatin microdose formulation consisting of 1% of the therapeutic dose of oxaliplatin and 2 × 106 dpm/kg of patient's body weight of labeled drug. We were able to correlate oxaliplatin–DNA adduct levels with clinical response in 5 of 6 patients. One patient with the highest adduct levels after 4 hours had an almost complete remission response, whereas the patient with the lowest adduct level did respond to FOLFOX, but to a much lesser degree. This early trend associating oxaliplatin–DNA adduct level with colorectal cancer patient response is encouraging. However, more patients need to be accrued to determine whether adduct levels are dependent upon tissue type and are predictive of clinical response.

Beyond the selection of oxaliplatin-based versus irinotecan-based initial therapy for palliative treatment of advanced colorectal cancer, there are multiple clinical scenarios where the application of diagnostic microdosing with oxaliplatin has the potential improve therapeutic decision making. For example, an understanding of the potential response to oxaliplatin-based therapy could improve the selection of treatment in patients with colorectal cancer liver metastases being considered for treated with intensive regimens such as FOLFOXIRI plus bevacizumab (24). In addition, variation in oxaliplatin–DNA adduct formation could potentially be used to optimize the selection of an oxaliplatin-based regimen as compared with gemcitabine and albumin-bound paclitaxel among patients with pancreatic cancer. Future investigation is required to translate our promising feasibility results to clinical utility in these settings. AMS analysis via conversion to graphite and quantification of the ratio of 14C to total carbon on specialized instrumentation is currently cost prohibitive and low throughput compared with the potential clinical need for many thousands of samples per year. However, the throughput and technology are improving with the recent advent of gas ionization sources and even laser-based systems that may render the diagnostic microdosing approach accessible to many laboratories and even hospitals settings.

In conclusion, we developed a highly sensitive AMS-based assay that can possibly identify cellular sensitivity to platinum-based drugs prior to toxic treatment. On the basis of this pilot study, a diagnostics feasibility clinical trial is currently in planning with enough power for statistical analysis to determine whether oxaliplatin–DNA monoadduct levels correlate with chemotherapy sensitivity, which would precede a subsequent study aimed at adjusting patient treatment planning based on the resulting data.

T.J. Semrad is a consultant/advisory board member for Celgene and Genentech. G. Cimino has ownership interest (including patents) in AMD. M. Cho has received speakers bureau honoraria from Taiho and is a consultant/advisory board member for Amgen and Astella. E.J. Kim is a consultant/advisory board member for Vicus, Armo, BluePath Solutions, Guidepoint, and Guardant Health. P.T. Henderson is a founder at Accelerated Medical Diagnostics and has ownership interest (including patents). No potential conflicts of interest were disclosed by the other authors.

Conception and design: M. Zimmermann, T.J. Semrad, G. Cimino, K.W. Turteltaub, C.-x. Pan, P.T. Henderson

Development of methodology: T. Li, C.-Y. Wu, A. Yu, K.W. Turteltaub, C.-x. Pan, P.T. Henderson

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Zimmermann, T. Li, T.J. Semrad, C.-Y. Wu, A. Yu, M. Malfatti, K. Haack, K.W. Turteltaub, C.-x. Pan, M. Cho, E.J. Kim

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Zimmermann, T. Li, T.J. Semrad, C.-Y. Wu, A. Yu, G. Cimino, K. Haack, C.-x. Pan, M. Cho, E.J. Kim, P.T. Henderson

Writing, review, and/or revision of the manuscript: M. Zimmermann, T.J. Semrad, A. Yu, G. Cimino, M. Malfatti, K.W. Turteltaub, C.-x. Pan, M. Cho, E.J. Kim, P.T. Henderson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Zimmermann, T.J. Semrad, A. Yu, P.T. Henderson

Study supervision: M. Zimmermann, T.J. Semrad, A. Yu, K.W. Turteltaub, C.-x. Pan, P.T. Henderson

SBIR contracts to Accelerated Medical Diagnostics Phase I HHSN261201000133C (to P.T. Henderson), phase II HHSN261201200048C (to P.T. Henderson), LLNL grants LDRD 08-LW-100 (to P.T. Henderson and M. Malfatti), NIH/NIGMS 2P41GM103483-16 (to K.W. Turteltaub), the Knapp Family Fund (to P.T. Henderson). This work was performed in part under the auspices of the U.S. DOE by LLNL under Contract DE-AC52-07NA27344 and supported by the National Institute of Health General Medical Sciences (2P41GM103483-16). The Research Resource is supported by the NIH, National Center for Research Resources, Biomedical Technology Program grant P41 RR13461. The UC Davis Comprehensive Cancer Center is supported by Cancer Center Support Grant P30CA093373 from the NCI. We are grateful to the patients and their families for enabling this work. This study was also supported by the UC Davis Comprehensive Cancer Center's Gastrointestinal Malignancies Innovation Group.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Marley
AR
,
Nan
H
. 
Epidemiology of colorectal cancer
.
Int J Mol Epidemiol Genet
2016
;
7
:
105
14
.
2.
Bahrami
A
,
Amerizadeh
F
,
Hassanian
SM
,
ShahidSales
S
,
Khazaei
M
,
Maftouh
M
, et al
Genetic variants as potential predictive biomarkers in advanced colorectal cancer patients treated with oxaliplatin-based chemotherapy
.
J Cell Physiol
2018
;
233
:
2193
201
.
3.
Tonini
G
,
Imperatori
M
,
Vincenzi
B
,
Frezza
AM
,
Santini
D
. 
Rechallenge therapy and treatment holiday: different strategies in management of metastatic colorectal cancer
.
J Exp Clin Cancer Res
2013
;
32
:
92
.
4.
Wang
S
,
Zhang
H
,
Scharadin
TM
,
Zimmermann
M
,
Hu
B
,
Pan
AW
, et al
Molecular dissection of induced platinum resistance through functional and gene expression analysis in a cell culture model of bladder cancer
.
PLoS One
2016
;
11
:
e0146256
.
5.
Stornetta
A
,
Zimmermann
M
,
Cimino
GD
,
Henderson
PT
,
Sturla
SJ
. 
DNA adducts from anticancer drugs as candidate predictive markers for precision medicine
.
Chem Res Toxicol
2016
;
30
:
388
409
.
6.
Pieck
AC
,
Drescher
A
,
Wiesmann
KG
,
Messerschmidt
J
,
Weber
G
,
Strumberg
D
, et al
Oxaliplatin-DNA adduct formation in white blood cells of cancer patients
.
Br J Cancer
2008
;
98
:
1959
.
7.
Arango
D
,
Wilson
AJ
,
Shi
Q
,
Corner
GA
,
Arañes
MJ
,
Nicholas
C
, et al
Molecular mechanisms of action and prediction of response to oxaliplatin in colorectal cancer cells
.
Br J Cancer
2004
;
91
:
1931
46
.
8.
Zimmermann
M
,
Wang
S-S
,
Zhang
H
,
Lin
T-y
,
Malfatti
M
,
Haack
K
, et al
Microdose-induced drug–DNA adducts as biomarkers of chemotherapy resistance in humans and mice
.
Mol Cancer Ther
2017
;
16
:
376
87
.
9.
Wang
SS
,
Zimmermann
M
,
Zhang
H
,
Lin
TY
,
Malfatti
M
,
Haack
K
, et al
A diagnostic microdosing approach to investigate platinum sensitivity in non-small cell lung cancer
.
Int J Cancer
2017
;
141
:
604
13
.
10.
Cimino
GD
,
Pan
CX
,
Henderson
PT
. 
Personalized medicine for targeted and platinum-based chemotherapy of lung and bladder cancer
.
Bioanalysis
2013
;
5
:
369
.
11.
Garner
RC.
Practical experience of using human microdosing with AMS analysis to obtain early human drug metabolism and PK data
.
Bioanalysis
2010
;
2
:
429
.
12.
Mosmann
T.
Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays
.
J Immunol Methods
1983
;
65
:
55
63
.
13.
Henderson
PT
,
Li
T
,
He
M
,
Zhang
H
,
Malfatti
M
,
Gandara
D
, et al
A microdosing approach for characterizing formation and repair of carboplatin-DNA monoadducts and chemoresistance
.
Int J Cancer
2011
;
129
:
1425
34
.
14.
Lévi
F
,
Metzger
G
,
Massari
C
,
Milano
G
. 
Oxaliplatin
.
Clin Pharmacokinet
2000
;
38
:
1
21
.
15.
Ognibene
TJ
,
Bench
G
,
Vogel
JS
,
Peaslee
GF
,
Murov
S
. 
A high-throughput method for the conversion of CO2 obtained from biochemical samples to graphite in septa-sealed vials for quantification of 14C via accelerator mass spectrometry
.
Anal Chem
2003
;
75
:
2192
6
.
16.
The 2007 Recommendations of the international commission on radiological protection: ICRP Publication 103
.
Ann ICRP
2007
;
37
:
1
332
.
17.
Brenner
DJ
,
Hall
EJ.
Computed tomography — an increasing source of radiation exposure
.
N Engl J Med
2007
;
357
:
2277
84
.
18.
Parker
RJ
,
Gill
I
,
Tarone
R
,
Vionnet
JA
,
Grunberg
S
,
Muggia
FM
, et al
Platinum—DNA damage in leukocyte DNA of patients receiving carboplatin and cisplatin chemotherapy, measured by atomic absorption spectrometry
.
Carcinogenesis
1991
;
12
:
1253
8
.
19.
Schellens
JH
,
Ma
J
,
Planting
AS
,
van der Burg
ME
,
van Meerten
E
,
de Boer-Dennert
M
, et al
Relationship between the exposure to cisplatin, DNA-adduct formation in leucocytes and tumour response in patients with solid tumours
.
Br J Cancer
1996
;
73
:
1569
75
.
20.
Fichtinger-Schepman
AMJ
,
van der Velde-Visser
SD
,
van Dijk-Knijnenburg
HCM
,
van Oosterom
AT
,
Baan
RA
,
Berends
F
. 
Kinetics of the formation and removal of cisplatin-DNA adducts in blood cells and tumor tissue of cancer patients receiving chemotherapy: comparison with in vitro adduct formation
.
Cancer Res
1990
;
50
:
7887
94
.
21.
Reed
E
,
Ozols
RF
,
Tarone
R
,
Yuspa
SH
,
Poirier
MC
. 
Platinum-DNA adducts in leukocyte DNA correlate with disease response in ovarian cancer patients receiving platinum-based chemotherapy
.
Proc Natl Acad Sci U S A
1987
;
84
:
5024
8
.
22.
Reed
E
,
Parker
RJ
,
Gill
I
,
Bicher
A
,
Dabholkar
M
,
Vionnet
JA
, et al
Platinum-DNA adduct in leukocyte DNA of a cohort of 49 patients with 24 different types of malignancies
.
Cancer Res
1993
;
53
:
3694
9
.
23.
Scharadin
TM
,
Zhang
H
,
Zimmermann
M
,
Wang
S
,
Malfatti
MA
,
Cimino
GD
, et al
Diagnostic microdosing approach to study gemcitabine resistance
.
Chem Res Toxicol
2016
;
29
:
1843
8
.
24.
Tomasello
G
,
Petrelli
F
,
Ghidini
M
,
Russo
A
,
Passalacqua
R
,
Barni
S
. 
FOLFOXIRI plus bevacizumab as conversion therapy for patients with initially unresectable metastatic colorectal cancer: a systematic review and pooled analysis
.
JAMA Oncol
2017
;
3
:
e170278
.