Background: Platinating agents are among the most commonly used cytotoxic drugs worldwide. It is recognized that Pt concentration can remain significantly increased in serum up to 20 years after completion of chemotherapy, with levels related to late treatment effects.
Methods: A Freedom EVO® Tecan liquid handler was used for aliquoting 50 μL serum at 10-fold dilution into 96-well plates. The Teledyne MVX-7100 low-volume autosampler was used for sample introduction into an Agilent 7900 inductively coupled plasma mass spectrometry. There was <1.2 min needed between injections. Time to completion for a maximum batch size using two 96-well plates was approximately 3.5 h, including preparation and analysis.
Results: Imprecision was <15%, and the limit of quantification was set at 5 ng/L based on imprecision of 18.3%. Bias based on fortified samples ranged from 0% to −4.3% within the analytical measurement range of 5–10 000 ng/L. The nonparametric reference interval for platinum in serum using 147 residual clinical samples was determined to be 8–47 ng/L. Serum platinum concentrations in 675 enrolled patients having an average time since chemotherapy completion of 6.4 (± 5.5 years) ranged from 20.1 to 8252.4 ng/L. Among all patients, 633 (94%) had serum concentrations exceeding 47 ng/L, and 42 (6%) had serum platinum concentrations between 8 and 47 ng/L.
Conclusions: This method used an automated liquid handler, a novel 96-well autosampler and 50 μL patient serum to quantify platinum levels. The method was successfully validated according to current clinical guidelines for laboratory developed tests.
Patients undergoing platinum-based chemotherapy will benefit from the information presented here. Evidence presented on residual serum platinum measured by inductively coupled plasma mass spectrometry (ICPMS) will allow better characterization of platinum-associated outcomes. Knowledge in the field of ICPMS as well as cancer chemotherapy will be advanced by the information presented here.
Platinating agents are among the most commonly used cytotoxic drugs worldwide. Platinum-based chemotherapy, however, is associated with a number of acute and long-term adverse effects such as bilateral hearing loss, tinnitus, peripheral neuropathy, Raynaud syndrome, infertility, and renal insufficiency, as well as late-onset sequelae such as hypertension, metabolic syndrome, dyslipidemia, diabetes, cardiovascular disease, subclinical hypogonadism, and secondary cancers (1).
Although it is well established that a cumulative dose of platinum administered intravenously is associated with a number of these adverse effects, it is also recognized that in European populations, platinum concentrations remain significantly increased in serum years after chemotherapy completion (2–4). Moreover, Brouwers et al. (5) showed that these circulating platinating agents remain partly reactive. Thus, an important question is the extent to which residual platinum in the serum might influence the risk of developing late effects (1). Sprauten et al. (3) recently documented a significant relationship between increasing concentration of residual serum platinum and the severity of neurotoxicity after adjusting for cumulative dose. Thus, it becomes increasingly apparent that quantification of long-term serum platinum concentrations in patients treated with platinum-based chemotherapy may help identify individuals at highest risk of late effects as well as elucidate possible mechanisms (4). Additionally, the extent to which persistent serum platinum concentrations might affect the actions of essential trace elements (e.g., calcium, magnesium) or result in chronic endothelial activation and vascular damage has also not been systematically or comprehensively addressed (1).
Previously published methods for high-sensitivity Pt measurements have used lengthy integration times or manual preparation using conventional autosamplers (5, 6). Here, we report the development and application of a robust inductively coupled plasma–based method using an automatic liquid handler for sample preparation and a low volume autosampler for sample introduction.
Materials and Methods
This project for development and measuring of Pt determinations in serum and its protocols was approved by the University of Utah Institutional Review Board (IRB #00007275). For the establishment of the reference intervals (RIs),7 all samples had been submitted in metal-free tubes for routine testing of trace elements and were determined to be within the established RI for Zn and Cu.
Residual Pt in serum of cancer survivors was measured in patients enrolled in the ongoing Platinum Study, which includes 8 cancer centers in the US and Canada (7–10). Eligibility criteria include men with a diagnosis of histologically or serologically confirmed germ cell tumor, age <50 years at diagnosis and 18+ years at study consent, treatment with cisplatin-based chemotherapy, and no subsequent salvage chemotherapy. Study procedures were approved by the Human Subjects Review Board at each institution. Serum platinum concentrations were measured in the first 675 enrolled patients, whose average age at clinical evaluation was 38.5 years (± 9.6 years). Cumulative cisplatin dose ranged from 198 to 800 mg/m2 [mean 362 (69) mg/m2]. Average time since chemotherapy completion was 6.4 years (± 5.5 years) (range 0.4–29.9 years).
Reference interval study
To establish the RI, 147 available clinical samples were selected based on the following criteria: (a) submitted for Zn or Cu serum testing in a certified trace element free tube, (b) results within the established RIs for Zn or Cu where applicable, (c) age ≥18 years, and (d) no previous cisplatin-based chemotherapy. A total of 63 females and 84 males were included with a nonparametric RI for both males and females age ≥18 years set at 8–47 ng/L.
A maximum batch size of 177 patient samples included 5 matrix-matched standards and 5 matrix-matched QC samples (run in duplicate, 1 set per plate) for a total of 192 injections from 2 × 96-well plates. Standards were prepared in synthetic serum (UTAK Laboratories) to reduce the contribution of endogenous Pt in the assay. For calibration standards, synthetic serum was fortified with 20, 100, 500, 2500, and 10 000 ng/L (to convert to nmol/L divide by 195.084) using a 1000 μg/mL stock [certified value 995 (5) μg/mL, purity 99.9952%] using 1% HNO3 for intermediate stock preparation. For QC materials, residual human serum was fortified to equal 75, 200, 1750, 4000, and 9500 ng/L using a certified stock of Pt (Inorganic Ventures). About 50 μL of sample was diluted in 1% HNO3 containing 1 μg/L Ir as the internal standard used for normalization. To equate cps of the internal standard to the 500 ng/L Pt calibrator, a 1 μg/L concentration of the internal standard solution was used.
An Agilent 7700x was used for analysis of the prepared samples. The peristaltic pump was operated at a flow rate of 0.3 mL/min for spray chamber evacuation only. A daily custom lens semi-autotune was conducted in addition to the initial hardware tune. Nickel sample and skimmer cones were used to reduce Pt background from conventional Pt-containing cones. Background Pt was verified as <25 cps each day of operation and was comparable to the cps of m/z 220. A one point peak pattern with 25 sweeps, 3 replicates, and 6 s of integration for 195Pt in standard and He mode were used. An integration time of 0.1s was used for Ir in both modes. Initially, a comparison of the ratio of 194Pt to 195Pt to the theoretical ratio for each sample was used to monitor for polyatomic and isobaric interferences of which none were noted. No further interference studies were performed, since it is well established and in our experience that hemoglobin, lipemia, and icterus do not significantly interfere with inductively coupled plasma mass spectrometry (ICPMS)-based methods due to the hard ionization of the plasma and the dilution of the sample for preparation (11). Further, it is highly unlikely that any of these classic indices would cause false measurements of Pt itself, but rather potential redistribution is the cited concern for trace element analysis with indices such as hemolysis (12). Quantitative data were based on standard mode measurement of 195Pt due to consistently lower relative standard deviations (CVs) when compared to He mode. Instrument parameters for a representative analytical run are provided in Table 1 in the Data Supplement that accompanies the online version of this article at http://www.jalm.org/content/vol1/issue2.
The MVX-7100 low-volume autosampler was used for sample introduction into the ICPMS equipped with a 1.5-mL sample syringe and 2.0-mL injection syringe. Before each batch, the sample loop and injection loop were rinsed using a sequence of 1.25 mL isopropyl alcohol, 1.25 mL 1% NH4OH, and 1.25 mL 1% HNO3 before 3 full syringe primes of a makeup fluid containing 0.5% HNO3 with 0.05% Triton X-100. A 500-μL sample loop was used to inject 450 μL of sample with airgaps at the front and the back to isolate the sample from the makeup solution and prevent mixing. The injection syringe was operated at 6.6 μL/s to produce an approximately 1-min integration time. The inclusion of He mode analysis with a 5-s stabilization time and the need to have a constant signal for spectrum mode analysis reduced the effective integration time to 45 s. Injection to injection time was <1.5 min.
A Freedom EVO® Tecan liquid handler (Tecan Group) was used for sample aliquoting and dilution into 96-well plates. Permanent tips were used to reduce the possibility of contamination from disposable tips with liquid sensing to ensure proper sample aspiration. Prepared 96-well plates were agitated for 60 s using the onboard plate shaker.
A standard addition calibration method was used with the origin ignored (not forced through zero) and 1/x weighting. No reagent blank or calibrator blank was used for standard curve generation. Data were analyzed and tabulated in Microsoft Excel (Microsoft Corporation), with figures and tables prepared using Excel and/or Adobe Illustrator (Adobe).
Evaluation of assay performance
We evaluated intraassay imprecision by running 20 replicates of 350 and 2000 ng/L in a single analytical run and interassay imprecision at the same concentrations using 4 separate analytical runs with 5 replicates over a span of 2 months for a total of 20 replicates. The limit of quantification (LOQ) was determined using a 5 ng/L serum pool across 4 separate analytical runs with 5 replicates for run over a span of 1 month. Because no alternative clinical assays existed for the determination of Pt in serum at low ng/L concentrations, verification of accuracy was conducted using fortification of serum pools from a certified Pt standard and determination of bias compared to the nonfortified pool. Carryover was assessed using a sequence of high, high, low1, and low2 samples with the percent carryover estimated to be [(low1 – low2)/high] * 100. Run stability was assessed by determining cps percent recovery of Ir for each sample in comparison to the cps for Ir in the lowest calibrator. Linearity was assessed using fortified serum pools from 20 000 to 1 ng/L with the analytical measurement range (AMR) set at 5–10 000 ng/L. Because this assay was developed for low-concentration Pt assessment, no dilution protocol was explored to extend the reportable range above 10 000 ng/L. A nonparametric RI was established using 147 residual serum samples submitted for Zn or Cu testing in trace element–free transport tubes. Samples for the RI establishment were from patients 18 years of age or older and having Zn and/or Cu results within the respective RIs (60–120 μg/dL, Zn; 70–140 μg/dL, Cu). Stability was assessed by repeat measurements of patient samples stored refrigerated for 6 months between replicate measures.
Method validation results are presented in Table 1, Table 2, Table 3 and Table 4. Intra- and interassay imprecision were <15% at 350 and 1950 ng/L. The LOQ was set at 5 ng/L based on an imprecision of 18.3% with a maximum tolerable imprecision limit of 20% equal to the functional sensitivity of the assay (13). Percent bias using fortified serum samples ranged from 0.6% to 5.4% for concentrations within the AMR of 5–10 000 ng/L. Carryover was determined to be 0.8% and was monitored for each analytical batch using the theoretical carryover to determine potential impact on any sampling immediately following an increased concentration ≥625 ng/L (i.e., the concentration where 0.8% carryover into a negative sample would equal 5 ng/L). Monitoring of the Ir cps for each injection of the maximum batch size analyzed during validation of 177 samples (5 calibration standards, 5 QC, and 86 samples plate 1; 5 QC and 91 patient samples plate 2) demonstrated acceptable variation of ±15% from the median with a maximum batch size using two 96-well plates of 192 samples in approximately 3.5 h including preparation and analysis. The RI for Pt, defined as the central 95% using the 2.5 and 97.5 percentiles, was established using 147 residual clinical samples and determined to be 8–47 ng/L (median, 12.4 ng/L; Q1, 9.6 ng/L; Q3, 14.6 ng/L) consistent with previous reports of low-concentration Pt detectable in healthy controls (5).
Residual serum Pt after cisplatin-based chemotherapy in cancer patients
Serum Pt concentrations ranged from 20.1 to 8252.4 ng/L [mean 601.2 (764.5) ng/L]. A total of 42 patients (6%) had serum Pt concentrations between 8 and 47 ng/L, and 633 patients (94%) had serum concentrations >47 ng/L (Fig. 1). Because neither the time since completion of chemotherapy nor serum platinum concentration was normally distributed (Kolmogorov–Smirnov test P value < 0.01), the Spearman correlation was used to assess the correlation between serum platinum concentration and time since completion of chemotherapy (correlation coefficient = −0.93, P < 0.0001). Time interval since completion of chemotherapy was further divided into <5 years, 5–10 years, 10–15 years, and >15 years (Table 5). As expected, with increasing time since completion of chemotherapy, the number of patients with serum platinum concentrations above the reference range decreased. Nonetheless, even 15 years after completion of chemotherapy, almost half of the patients (44%) had serum platinum concentrations that exceeded the reference range. The AMR was determined to be 5–10 000 ng/L. Ten samples that spanned the established AMR were analyzed before and after storage at 4 °C for 6 months and repeated within 10% of the original value (range 3.2–9.7%).
Previously published methods for quantifying low Pt concentrations in serum represented high-quality research assays but were difficult to implement for routine use in the clinical laboratory (4–6, 14). Original methods were based on the measurement of Pt in plasma ultrafiltrates and used lengthy integration times to achieve the needed sensitivity (6). Subsequent methods for serum also required lengthy integration times and did not include specific details regarding assay performance characteristics in a substantially more complicated matrix (5). A more recent method used magnetic sector mass spectrometry, which is not readily applicable to high-volume testing in a routine clinical laboratory environment (15). These assays are not easily implemented in the clinical laboratory or amenable for high-throughput clinical studies due to lengthy integration times, high complexity platforms, complicated dilution protocols, or validation of performance differing from current CLIA requirements and recommendations. The current assay demonstrated acceptable performance for all measured parameters, used automated liquid handling for sample preparation, and used a low-volume autosampler in a 96-well format. Use of the low-volume autosampler allowed for reduced clinical sample requirements (50 μL for a single analysis) and a total prepared sample volume of 500 μL, making the presented method ideal for biobank studies using limited samples and routine use in a clinical laboratory. The established RI in healthy controls identified Pt in all samples using the LOQ of 5 ng/L and is consistent with previous reports (5).
To our knowledge, to date only Sprauten et al. (3) and Boer et al. (4) have addressed the extent to which serum Pt concentrations are correlated with cisplatin-associated toxicities. Both studies documented a significant relationship between increasing concentrations of residual serum Pt and the severity of neurotoxicity in testicular cancer survivors treated with cisplatin-based chemotherapy, even when analyses were adjusted for initial cumulative cisplatin dose (3, 4). The long-term persistence of increased serum Pt concentrations has been known for several decades, with investigations documenting amounts up to 1000-fold greater than in unexposed controls (2, 5, 16, 17). Because circulating Pt remains partly reactive (5), one of the key research recommendations at an international consensus conference (1) involved elucidating the relationship between decades-long exposure to low serum Pt concentrations and the development of biological effects. Moreover, additional study of the extent to which serum Pt might affect the actions of essential trace elements (e.g., calcium, copper, magnesium, iron, and zinc) was also recommended. It is well established that cisplatin administration can be acutely associated with hypomagnesemia (18), although data describing its long-term persistence are conflicting (19–22). Hypomagnesemia has been reported for more than 6 years after chemotherapy in some series (21, 22), but not others (19, 20, 22).
The results of Sprauten et al. (3) and Boer et al. (4) suggest that total serum Pt concentration may serve as a useful biomarker of late toxicity and body burden. Long-term serum Pt concentrations may represent its release from various body reservoirs during tissue remodeling [reviewed in (1)]. Pt-DNA adducts have been measured in numerous tissues (e.g., kidney, brain) after Pt-based chemotherapy (17, 23), and further study of the long-term deposition of Pt (sites and reactivity), serum concentrations, and correlation with late effects has also been recommended (1). The biokinetic properties of Pt mimic those of other toxic metals (e.g., mercury and chromium) in which serum and urine concentrations also correlate with adverse outcomes (24). Four patients in our present data set had visibly increased Pt results >5 years posttreatment in comparison to the majority Fig. 1. In a study by Boer et al. (4), the pharmacokinetic model for Pt elimination was found to vary with renal function and renal impairment significantly decreasing elimination. In addition, our data set does not rule out the possibility of an environmental exposure to Pt through routes such as Pt salts used in the preparation of sulfuric acid and petroleum or platinum used in automobile catalytic converters (25).
The acute cardiovascular effects of cisplatin, including its thrombogenicity, have recently been reviewed (26) and may result in significantly increased risk of death due to cardiovascular disease (CVD) including stroke. However, a growing body of literature summarized by Fung et al. (26) also points to the significantly increased incidence of CVD in long-term testicular cancer survivors (TCS). The extent to which circulating Pt concentrations may contribute to this outcome has not yet been addressed to our knowledge, but would also have to take into account general CVD risk factors, such as tobacco use, level of physical activity, obesity, and smoking [reviewed in (1)].
Fung et al. also recently reported a significantly increased risk of solid tumors after chemotherapy, most likely containing cisplatin, among 6013 TCS in the population-based Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute (8). Prior studies have established strong dose–response relationships between cumulative cisplatin dose and secondary leukemia, and analytic investigations of solid tumor risk are needed (27, 28). Fung et al. showed that 3-fold increased risks of kidney cancer follow chemotherapy for testicular cancer (29), and in this regard, it is noteworthy that increased urine concentrations of Pt have also been observed for several decades after platinum administration [reviewed in (1)].
Although much of the work to date with regard to the long-term effect of increased serum Pt concentrations have been restricted to TCS, it is likely that in the future, other patient populations will also be studied, since the Pt compounds remain the most commonly used group of cytotoxic drugs worldwide. In particular, cisplatin is used to effectively treat many types of childhood cancer, with survivors remaining at risk for a lifetime for the late effects of these and other types of successful treatments.
The authors thank all members of the Trace and Toxic Elements laboratory at ARUP Laboratories and the ARUP Institute for Clinical and Experimental Pathology for financial support. F.G. Strathmann thanks Emmett Soffey for helpful discussions regarding Pt assessment by ICPMS.
↵7 Nonstandard abbreviations:
- reference interval
- inductively coupled plasma mass spectrometry
- limit of quantification
- analytical measurement range
- cardiovascular disease
- testicular cancer survivors
- Surveillance, Epidemiology, and End Results.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Employment or Leadership: S. Moody, D. Clarke, Teledyne, Inc.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: F.G. Strathmann, C.L. Law, ARUP Institute for Clinical and Experimental Pathology; L.B. Travis, the National Cancer Institute, grant 1R01 CA157823-01A1.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
- Received April 24, 2016.
- Accepted June 20, 2016.
- © 2016 American Association for Clinical Chemistry