Prostate cancer is the commonest noncutaneous malignancy in men worldwide, and its incidence is expected to double by 2030, in part due to an aging global population.1,2 3–5 6
Anatomical T2-weighted imaging is limited by the nonspecific nature of low signal intensity findings in the peripheral zone (PZ).7 8,9 10–12
Standard clinical 1 H-MRI provides information about the distribution of water molecules, but MRI can be used to detect several nuclei other than hydrogen, thus allowing for other physical and metabolic processes to be studied in the human body. One such example is sodium (23 Na), which produces the second highest signal on MRI from the nuclei detectable in biological tissues. However, the MRI sensitivity for sodium is approximately 9.2% of protons, and the sodium concentration in vivo is approximately 2000 times lower than that of protons, resulting in signal-to-noise ratios (SNRs) up to 20,000 times lower than 1 H-MRI.13 23 Na-MRI acquisition in a clinically acceptable time frame (10–15 minutes) and with an isotropic resolution of a few millimeters.14 23 Na-MRI has become a viable addition to existing 1 H-MRI protocols.
Changes in tissue sodium occur in a wide range of disease processes including inflammation, ischemia, and in several tumors, with the measurements being sensitive to changes in both energetic status and the integrity of cell membranes. Therefore, studying the sodium concentrations in prostate cancer could provide complementary information to traditional 1 H-MRI. Few 23 Na-MRI studies on human prostate are published, with only 3 studies assessing healthy volunteers and a further study investigating 3 patients with biopsy-proven prostate cancer, without quantifying tumor sodium concentration.15–18
MATERIALS AND METHODS
The local institutional review board and ethics committee granted approval for this prospective study (CUH/16/EE/0205), with all participants signing written informed consent. Fifteen patients with biopsy-proven, MRI visible (PI-RADS score 4 or 5 lesions on prior diagnostic prostate MRI), intermediate- or high-risk prostate cancer underwent a dedicated research sodium MRI between May 2016 and August 2017, before treatment with radical prostatectomy. Exclusion criteria were previous treatment for prostate cancer, a clinical contraindication to MRI, or renal impairment (glomerular filtration rate <30 mL/min).
Magnetic Resonance Imaging
All phantom experiments and clinical imaging were performed on a clinical 3 T system (GE MR750; GE Healthcare, Waukesha, WI) using a dedicated multinuclear clamshell transmit coil with a large bandwidth (GE Healthcare, Waukesha, WI) and a bespoke dual-tuned 1 H/23 Na endorectal receive coil.19
Sodium imaging was performed using a 3-dimensional cones20,21 1 H fast-relaxation fast spin echo (FRFSE) T2-weighted anatomical images were acquired using the following parameters: TE, 105 milliseconds; TR, 2500 milliseconds; FOV, 30 cm; resolution, 0.5 × 0.5 × 4 mm.
For intracellular-weighted imaging, the sodium sequence was repeated with the addition of a 5.6-millisecond adiabatic inversion pulse with inversion time (TI) of 30 milliseconds, TR of 120–210 milliseconds, resolution and FOV as before, 10 NEX, 5390 total readouts, and 11 to 19 minutes' acquisition duration. The variation in TR was due to restrictions imposed by the Specific Absorption Rate (SAR) limit and varied between patients due to differences in weight, which ranged from 65 to 109 kg (mean weight, 83.5 ± 12.7 kg). For the inversion recovery sodium imaging, TR and therefore total acquisition duration was varied due to SAR limitations. Sodium images were corrected for receive sensitivity by dividing the images by a heavily smoothed duplicate image (Gaussian kernel, sigma = 522 Fig. 1 ). A phantom replacement method was used to create a calibration curve against 5 NaCl in 4% agar phantoms varying between 7 and 160 mM/L NaCl, and this was applied to the in vivo data to create quantitative sodium maps.23
FIGURE 1: Sensitivity-corrected images. Proton T2-weighted image (A and B) and TSC images (C and D). The uncorrected images (A and C) were adjusted for receive sensitivity by dividing the images by a heavily smoothed duplicate to create the corrected images (B and D). Images displayed at full dynamic range.
Histopathology Assessment
Prostatectomy specimens were fixed in formalin and oriented by the location of the seminal vesicles, posterior surface of the prostate, and by the position of the urethra. The apical cone was amputated and sliced into 4-mm sections from left to right; the remaining gland was cut transversely into 5-mm whole-mount parallel slices in the horizontal plane from inferior to superior. Representative 5-μm microtome slides were processed from each 5-mm whole-mount slices for histopathological analysis. Tumor was outlined on hematoxylin and eosin–stained sections from each slice by an experienced uropathologist specializing in prostate cancer.
Correlation of Histopathology to Imaging
A radiologist with 7 years of experience reporting clinical prostate MRI used the histopathological tumor maps in conjunction with the available diagnostic MRI data sets to outline peripheral zone (PZ), transition zone (TZ), and tumor. The tumor location from histopathological slides was used to manually delineate an ROI on the T2-weighted anatomical image acquired using the endorectal coil. Regions of interest of a minimum volume of 0.5 cm3 were additionally drawn on at least 3 consecutive slices for normal TZ and PZ, with care taken to avoid partial volume effects from the urethra, bladder, seminal vesicles, extraprostatic tissue, and at the zonal interfaces. Separately, to assess whether the reduced SNR in the anterior gland affected estimates of sodium concentration, ROIs in all patients were drawn in the anterior half and posterior half of the TZ. Regions of interest based on these outlines were then copied to the colocalized sodium images using OsiriX 8.5.1 (Pixmeo SARL, Bernex, Switzerland).
Statistics
A balanced 1-way analysis of variance was performed to compare the results from each patient for normal PZ, normal TZ, and tumor ROIs, using Tukey test for multiple comparisons. A 2-sided Wilcoxon rank sum test was used to assess for differences in sodium values between tumor grades. A P value of less than 0.05 was considered statistically significant.
RESULTS
Magnetic resonance imaging was acquired without complications, and all patients proceeded to prostatectomy. The 15 participants had a median age of 59.5 years (mean, 60.7; range, 48–73 years) and a median prebiopsy serum PSA of 8.4 ng/mL (mean, 9.79; interquartile range, 6.125–11.6 ng/mL; Table 1 ). The average time between diagnostic prostate biopsy and the study MRI was 89.6 days (range, 43–179 days). Eight patients underwent MRI scanning on the day of surgery; the mean time from MRI to surgery was 5 days (median, 0; range, 0–31 days). The total sodium MRI time ranged from 18 to 26 minutes. The whole examination time including proton scans and coil insertion was on average of 37 minutes (range, 35–47 minutes).
TABLE 1: Patient Characteristics
Seventeen tumors were detected in the 15 patients imaged, with final pathology of Gleason score 3 + 3 (n = 1), 3 + 4 (n = 7), 3 + 5 (n = 2), 4 + 3 (n = 5), and 4 + 5 (n = 2). The mean tumor size at final pathology was 4.21 cm3 (range, 0.09–15.78 cm3 ). Two patients had significant artifact on the sodium imaging and had to be excluded (Gleason 3 + 4 and 4 + 5 tumors), and in 1 further patient, intracellular-weighted sodium data could not be acquired due to a software error (Gleason 3 + 4 tumor). Thus 15 tumors in 13 patients were analyzed for TSC and 14 tumors in 12 patients for intracellular sodium.
The mean TSC was significantly higher in the normal peripheral zone at 39.2 mmol/L (range, 34.2–42.5) than in the normal transition zone 32.9 mmol/L (range, 28.4–39.5; P < 0.001, Fig. 2 ). The mean TSC for all tumors was 43.1 mmol/L and for PZ tumors 45.0 mmol/L (range, 38.5–50.9; Table 2 , Fig. 3 ). Overall, 11/13 peripheral zone tumors had a higher TSC than the corresponding normal PZ on the contralateral side of the gland. The mean PZ tumor TSC was significantly higher than both the normal PZ TSC (P < 0.001) and the normal TZ TSC (P < 0.001; Fig. 4 ). Two of the 15 tumors were located in the TZ; their mean TSC was 30.6 and 30.8 mmol/L, which was slightly lower than the TSC in the adjacent normal TZ.
FIGURE 2: Sodium concentration in the normal PZ and TZ. A 58-year-old man, PSA level of 18.69 ng/mL with a pathologically proven Gleason score of 3 + 5 tumor in the left apical TZ (not shown). Regions of interest outline normal PZ (blue) and normal TZ (yellow). A, T2-weighted axial image. B, Map of tissue sodium concentration. C, Map of intracellular sodium concentration. The mean TSC and intracellular sodium concentrations were 40.6 and 10.2 mmol/L in the PZ, and 32.7 and 7.9 mmol/L in the TZ, respectively.
TABLE 2: Mean TSC in mml/L for Normal and Tumor Tissue
FIGURE 3: Sodium concentration in PZ tumor. A 63-year-old man, PSA level of 8.4 ng/mL with a 20-mm pathologically proven Gleason 4 + 3 tumor in the right mid PZ, with focal extracapsular extension. A, Diagnostic T2-weighted image with lesion in the right mid PZ (arrow), confirmed on whole-mount histology (B, outline). C, TSC map. D, TSC map overlaid on T2 image. The mean TSC and intracellular sodium concentrations in the lesion were 41.8 and 19.2 mmol/L, compared with 37.9 and 17.1 mmol/L in the normal PZ, respectively.
FIGURE 4: Box and whiskers plot of the tissue sodium concentration in normal and PZ tumor tissue. Top and bottom of boxes represent 25th and 75th percentiles of data, lines in boxes represent the median value, and bars represent the data within 1.5 times interquartile range; + denotes outliers. P values between groups as indicated. PZ indicates normal peripheral zone; TZ, normal transition zone. There were only 2 TZ tumors in the cohort; their values are denoted by + symbols.
The mean intracellular-weighted sodium was also significantly higher in the normal PZ (17.5 mmol/L; range, 10.2–21.4) than in the normal TZ (14.7 mmol/L; range, 7.9–18.0; P = 0.02). The mean intracellular-weighted sodium for all tumors was 18.7 mmol/L and for PZ tumors 19.9 mmol/L (range, 15.8–22.3). Overall 9/12 PZ tumors had higher intracellular sodium content than the adjacent normal PZ (borderline significant, P = 0.05), with PZ tumor tissue having a significantly higher intracellular sodium than normal TZ (P < 0.001; Fig. 5 ). Both TZ lesions had slightly higher intracellular sodium values than the normal TZ (Table 3 ).
FIGURE 5: Box and whiskers plot of intracellular-weighted sodium concentration in normal and PZ tumor tissue. Top and bottom of boxes represent 25th and 75th percentiles of data, lines in boxes represent the median value, and bars represent the data within 1.5 times the interquartile range; + denotes outliers. P values between groups as indicated. PZ indicates normal peripheral zone; TZ, normal transition zone. There were only 2 TZ tumors in the cohort; their values are denoted by + symbols.
TABLE 3: Mean Intracellular-Weighted Sodium Concentration in mmol/L for Normal and Tumor Tissue
Regional assessment of the differences in signal between the anterior and posterior gland showed a significant variation. The mean TSC value in the anterior TZ was 30.51 mmol/L (±1.67) compared with 33.48 mmol/L (±1.86) in the posterior TZ (P = 0.004). The mean intracellular sodium value was 14.32 mmol/L (±0.82) in the anterior TZ, compared with 15.53 mmol/L (±1.06) in the posterior TZ (P = 0.08).
The association between sodium concentration and tumor grade was also analyzed. Of the PZ tumors, 7 were Gleason 3 + 3 or 3 + 4 tumors, and 6 were Gleason ≥ 4 + 3. The mean TSC was lower in those with Gleason ≤ 3 + 4 (44.4 mmol/L) than in those with ≥ 4 + 3 (45.6 mmol/L), but the difference was not statistically significant (P = 0.19). Likewise, there was no significant difference in mean intracellular sodium between Gleason ≤ 3 + 4 (19.5 mmol/L) and Gleason ≥ 4 + 3 (20.2 mmol/L) tumors (P = 0.29).
DISCUSSION
This study demonstrates the ability of MRI to quantify TSC and intracellular sodium concentration in both tumors and normal prostate tissue. Tissue sodium concentration has previously been quantified in the normal PZ and TZ of both volunteers and patients. Here we developed sodium MRI at a higher spatial resolution than previously achieved and quantify tissue and intracellular sodium concentrations within prostate tumors for the first time.
The total TSC was significantly higher in tumors than in normal tissue, and the intracellular sodium was higher in tumors than in normal TZ. We found a significant difference between intracellular sodium in normal PZ and TZ (17.5 and 14.7 mmol/L). The values measured are within the expected range for other tissues, with intracellular sodium concentrations previously reported to be approximately 10 to 15 mmol/L.13,24,25 15 15,16,18
Physiologically, extracellular sodium concentration (140–150 mmol/L) is an order of magnitude higher than intracellular sodium concentration; however, the extracellular volume fraction (including vessels) accounts for only approximately 20% of total tissue volume.13,25 26–28 15
In healthy tissues, the large concentration gradient between intracellular and extracellular sodium is maintained by several transporters and pumps, the most important being Na+ /K+ -ATPase.29 + /K+ -pump to maintain the sodium gradient, or due to overactivity of the Na+ /K+ -pump to compensate for the low pH of the tumor microenvironment induced by hypoxia and increased lactate.13,30,31 32 33 34
We found a significantly higher TSC in PZ tumors compared with both normal PZ and normal TZ. This implies that the change in sodium concentration in tumors reflects more than a simple change in cellularity. Prostate tumors have a higher cell density than the normal PZ, typically seen as low ADC values. Given that intracellular sodium is significantly lower than extracellular sodium, cellular changes alone would lead to a reduction rather than an increase in TSC, by increasing the percentage contribution of intracellular sodium at the expense of a reduced extracellular space. Intracellular sodium in PZ tumors was higher than in normal PZ and TZ, with the difference being significant compared with the TZ and borderline significant compared with the PZ; this change in intracellular sodium could reflect changes in cellularity alone. Taken together, these results suggest that there is an increase in extracellular sodium concentration in PZ tumors. Interestingly, the 2 TZ tumors had lower TSCs compared with the PZ tumors, which may be partly explained by the drop-off in SNR with increasing distance from the coil: PZ tumors are unlikely to be affected due to their relative proximity to the receive coil. However, the lower sodium values in these TZ tumors relative to normal TZ implies that this is a genuine finding and indicates a different sodium environment in TZ and PZ tumors. Indeed, TZ tumors have morphological differences compared with PZ tumors, with well-differentiated glands of variable size and are lined by columnar cells with a clear cell histological pattern.35,36
There may be several potential applications for sodium MRI in prostate cancer, including localization and characterization of the primary tumor, monitoring treatment response to oncological agents, and detecting local recurrence following radical therapy. Furthermore, any treatment that causes cell death or loss of membrane integrity, such as the many focal ablative therapy techniques that are being trialed in prostate cancer, could potential be monitored with sodium MRI. The detection of sodium signal in vivo remains challenging and improvements in hardware and software and/or increasing magnet strength are needed to help improve SNR and resolution and reduce acquisition times and make this promising technique more clinically available. Importantly, there is increasing scientific interest in the acquisition of hyperpolarised carbon-13 MRI, particularly in the prostate37 13 C hardware and software will directly impact on the ability to image sodium and increase the availability of equipment, sequences, and MRI physics expertise.
Our study has some limitations. The total number of tumors was relatively small, and only 2 TZ tumors were included. The selection bias of patients undergoing prostatectomy while affording definitive histology limits the ability to differentiate grade type in this cohort, with an overrepresentation of Gleason 3 + 4 and 4 + 3 disease and underrepresentation of higher-grade and lower-grade Gleason 3 + 3 disease. Sodium imaging was performed after anatomical T2 imaging and theoretically interim bladder filling could affect prostate position and lead to mismatch between the sequences. Although some studies report that a large bladder volume can displace the prostate,38,39 40,41
In conclusion, we successfully quantified the tissue and intracellular sodium concentrations of prostate tumors in vivo, demonstrating a significant increase in TSC within PZ tumors. Sodium MRI may be developed as a new, noninvasive, functional imaging technique in the management of patients with prostate cancer.
ACKNOWLEDGMENTS
The authors acknowledge grant support from the Evelyn Trust UK and research support from Cancer Research UK, National Institute of Health Research Cambridge Biomedical Research Centre, Cancer Research UK, and the Engineering and Physical Sciences Research Council Imaging Centre in Cambridge and Manchester, as well as the Cambridge Experimental Cancer Medicine Centre.
REFERENCES
1. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017.
CA Cancer J Clin . 2017;67:7–30.
2. Maddams J, Utley M, Møller H. Projections of cancer prevalence in the United Kingdom, 2010–2040.
Br J Cancer . 2012;107:1195–1202.
3. de Rooij M, Hamoen EH, Fütterer JJ, et al. Accuracy of multiparametric MRI for prostate cancer detection: a meta-analysis.
AJR Am J Roentgenol . 2014;202:343–351.
4. Woo S, Suh CH, Kim SY, et al. Diagnostic performance of prostate imaging reporting and data system version 2 for detection of prostate cancer: a systematic review and diagnostic meta-analysis.
Eur Urol . 2017;11. pii: S0302-283830067–30062.
5. Turkbey B, Albert PS, Kurdziel K, et al. Imaging localized prostate cancer: current approaches and new developments.
AJR Am J Roentgenol . 2009;192:1471–1480.
6. Barrett T, Haider MA. The emerging role of MRI in prostate cancer active surveillance and ongoing challenges.
AJR Am J Roentgenol . 2017;208:131–139.
7. Turkbey B, Pinto PA, Mani H, et al. Prostate cancer: value of multiparametric MR imaging at 3 T for detection—histopathologic correlation.
Radiology . 2010;255:89–99.
8. Weinreb JC, Barentsz JO, Choyke PL, et al. PI-RADS Prostate Imaging - Reporting and Data System: 2015, Version 2.
Eur Urol . 2016;69:16–40.
9. Barrett T, Turkbey B, Choyke PL. PI-RADS version 2: what you need to know.
Clin Radiol . 2015;70:1165–1176.
10. Hambrock T, Somford DM, Huisman HJ, et al. Relationship between apparent diffusion coefficients at 3.0-T MR imaging and Gleason grade in peripheral zone prostate cancer.
Radiology . 2011;259:453–461.
11. Barrett T, Priest AN, Lawrence EM, et al. Ratio of tumor to normal prostate tissue apparent diffusion coefficient as a method for quantifying DWI of the prostate.
AJR Am J Roentgenol . 2015;205:W585–W593.
12. Donati OF, Mazaheri Y, Afaq A, et al. Prostate cancer aggressiveness: assessment with whole-lesion histogram analysis of the apparent diffusion coefficient.
Radiology . 2014;271:143–152.
13. Madelin G, Regatte RR. Biomedical applications of sodium MRI in vivo.
J Magn Reson Imaging . 2013;38:511–529.
14. Ouwerkerk R. Sodium MRI.
Methods Mol Biol . 2011;711:175–201.
15. Hausmann D, Konstandin S, Wetterling F, et al. Apparent diffusion coefficient and sodium concentration measurements in human prostate tissue via hydrogen-1 and sodium-23 magnetic resonance imaging in a clinical setting at 3 T.
Invest Radiol . 2012;47:677–682.
16. Farag A, Peterson JC, Szekeres T, et al. Unshielded asymmetric transmit-only and endorectal receive-only radiofrequency coil for (23) Na MRI of the prostate at 3 tesla.
J Magn Reson Imaging . 2015;42:436–445.
17. Bae KT, Kim J-H, Furlan A, et al. Proton and sodium MR imaging of in vivo human prostate using dual-tuned body and endorectal coils at 7 T.
Proc Intl Soc Mag Reson Med . 2010(18):2693.
18. Paschke NK, Hausmann D, Schad LR, et al. Multi-parametric
1 H/
23 Na clinical protocol of the prostate at 3 T using a double resonance coil.
Proc Intl Soc Mag Reson Med . 2017(25):1016.
19. Tropp J, Calderon P, Carvajal L, et al. An endorectal dual frequency
13 C-
1 H receive only probe for operation at 3.0 tesla.
Proc Intl Soc Mag Reson Med . 2006(14):2594.
20. Gurney PT, Hargreaves B, Nishimura DG. Design and analysis of a practical 3D cones trajectory.
Magn Reson Med . 2006;55:575–582.
21. Riemer F, Solanky BS, Stehning C, et al. Sodium ((23)Na) ultra-short echo time imaging in the human brain using a 3D-Cones trajectory.
MAGMA . 2014;27:35–46.
22. Christensen JD, Barrère BJ, Boada FE, et al. Quantitative tissue sodium concentration mapping of normal rat brain.
Magn Reson Med . 1996;36:83–89.
23. Axel L, Costantini J, Listerud J. Intensity correction surface-coil MR imaging.
Am J Roentgenol . 1987;148:418–420.
24. Conway EJ. Nature and significance of concentration relations of potassium and sodium ions in skeletal muscle.
Physiol Rev . 1957;37:84–132.
25. Madelin G, Kline R, Walvick R, et al. A method for estimating intracellular sodium concentration and extracellular volume fraction in brain in vivo using sodium magnetic resonance imaging.
Sci Rep . 2014;4:4763.
26. Wang XZ, Wang B, Gao ZQ, et al. Diffusion-weighted imaging of prostate cancer: correlation between apparent diffusion coefficient values and tumor proliferation.
J Magn Reson Imaging . 2009;29:1360–1366.
27. Epstein JI, Allsbrook WC Jr, Amin MB, et al. The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma.
Am J Surg Pathol . 2005;29:1228–1242.
28. Lawrence EM, Warren AY, Priest AN, et al. Evaluating prostate cancer using fractional tissue composition of radical prostatectomy specimens and pre-operative diffusional kurtosis magnetic resonance imaging.
PLoS One . 2016;11:e0159652.
29. Murphy E, Eisner DA. Regulation of intracellular and mitochondrial sodium in health and disease.
Circ Res . 2009;104:292–303.
30. Cameron IL, Smith NK, Pool TB, et al. Intracellular concentration of sodium and other elements as related to mitogenesis and oncogenesis in vivo.
Cancer Res . 1980;40:1493–1500.
31. Rotin D, Steele-Norwood D, Grinstein S, et al. Requirement of the Na+/H+ exchanger for tumor growth.
Cancer Res . 1989;49:205–221.
32. Jacobs MA, Ouwerkerk R, Wolff AC, et al. Monitoring of neoadjuvant chemotherapy using multiparametric,
23 Na sodium MR, and multimodality (PET/CT/MRI) imaging in locally advanced breast cancer.
Breast Cancer Res Treat . 2011;128:119–126.
33. Sharma R, Katz JK. Taxotere chemosensitivity evaluation in mice prostate tumor: validation and diagnostic accuracy of quantitative measurement of tumor characteristics by MRI, PET, and histology of mice tumor.
Technol Cancer Res Treat . 2008;7:175–185.
34. Jacobs MA, Ouwerkerk R, Kamel I, et al. Proton, diffusion-weighted imaging, and sodium (
23 Na) MRI of uterine leiomyomata after MR-guided high-intensity focused ultrasound: a preliminary study.
J Magn Reson Imaging . 2009;29:649–656.
35. Grignon DJ, Sakr WA. Zonal origin of prostatic adenocarcinoma: are there biologic differences between transition zone and peripheral zone adenocarcinomas of the prostate gland?
J Cell Biochem . 1994;19(Suppl):267–269.
36. McNeal JE, Redwine EA, Freiha FS, et al. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread.
Am J Surg Pathol . 1988;12:897–906.
37. Nelson SJ, Kurhanewicz J, Vigneron DB, et al. Metabolic imaging of patients with prostate cancer using hyperpolarized [1-
13 C]pyruvate.
Sci Transl Med . 2013;5:198ra108.
38. Crook JM, Raymond Y, Salhani D, et al. Prostate motion during standard radiotherapy as assessed by fiducial markers.
Radiother Oncol . 1995;37:35–42.
39. Ten Haken RK, Forman JD, Heimburger DK, et al. Treatment planning issues related to prostate movement in response to differential filling of the rectum and bladder.
Int J Radiat Oncol Biol Phys . 1991;20:1314–1324.
40. Pinkawa M, Asadpour B, Gagel B, et al. Prostate position variability and dose-volume histograms in radiotherapy for prostate cancer with full and empty bladder.
Int J Radiat Oncol Biol Phys . 2006;64:856–861.
41. Villeirs GM. Magnetic resonance assessment of prostate localization variability in intensity-modulated radiotherapy for prostate cancer.
Int J Radiat Oncol Biol Phys . 2004;60:1611–1621.
