Three-dimensional pulsed continuous arterial spin labeling and intravoxel incoherent motion imaging of nasopharyngeal carcinoma: correlations with Ki-67 proliferation status

Abstract

Background: Recurrence and distant metastasis are still the main problems affecting the long-term prognosis of nasopharyngeal carcinoma (NPC) patients, and may be related to the Ki-67 proliferation status. We therefore explored the potential correlation between Ki-67 proliferation status in NPC with the parameters derived from two imaging techniques: three-dimensional pulsed continuous arterial spin labeling (3D pCASL) and intravoxel incoherent motion (IVIM).

Methods: Thirty-six patients with pathologically confirmed NPC were included, and the Ki-67 labeling index (LI) was measured by immunohistochemistry. All patients underwent plain and contrast-enhanced magnetic resonance imaging (MRI), IVIM, and 3D pCASL examination. The mean, maximum, and minimum of blood flow (BF), minimum of apparent diffusion coefficient (ADC), pure diffusion coefficient (D), pseudodiffusion coefficient (D*), and perfusion fraction (f) parameters were all measured, and Spearman's correlation analysis was performed to evaluate the relationships between these parameters and the Ki-67 LI. According to the Ki-67 values, the patients were divided into two groups: high (>50%) and low (≤50%). The rank-sum test (Mann-Whitney U test) was then used to compare the differences in quantitative parameters between the high and low Ki-67 groups.

Results: Ki-67 LI was positively correlated with BF_mean and BF_max (r=0.415 and 0.425). D*_mean and D*_min did have positive correlation with Ki-67, but this was not significant (P=0.082 and 0.072). BF_max was significantly different between the high and low Ki-67 groups (P=0.028).

Conclusions: 3D pCASL and IVIM are noninvasive functional MR perfusion imaging techniques that can evaluate perfusion information and perfusion parameters. Our study suggests that 3D pCASL is more effective than IVIM for assessing the proliferation status of NPC, which is beneficial for evaluating the prognosis of patients. Furthermore, BF_max is the best biomarker for distinguishing high from low Ki-67 levels.

Keywords: Ki-67 antigen; Spin labels; diffusion magnetic resonance imaging; nasopharyngeal carcinoma (NPC).

Figure 1

Flowchart depicting the patient selection process.

Figure 2

Correlations between the quantitative MR parameters (ASL and IVIM) and the Ki-67 labeling index in NPC. (A) The BF_mean were positively correlated with Ki-67 and the correlation coefficients r were 0.415. (B) BF_max, were positively correlated with Ki-67 and the correlation coefficients r were 0.425 (P<0.05). (C) D*_mean were not significantly correlated with Ki-67, as the P values were 0.082. (D) D*_min were not significantly correlated with Ki-67, as the P values were 0.072, but a positively correlation with Ki-67 was nonetheless observed. (E) BF_min were not significantly correlated with Ki-67 (P>0.05). (F) ADC_mean were not significantly correlated with Ki-67 (P>0.05). NPC, nasopharyngeal carcinoma; ASL, arterial spin labeling; IVIM, intravoxel incoherent motion; BF, blood flow; BF_mean, mean BF; BF_max, maximum BF; BF_min, minimum BF; ADC, apparent diffusion coefficient; ADC_mean, mean ADC; D*, pseudo diffusion coefficient; D*_mean, mean D*; D*_min, the minimum ADC corresponding D*.

Figure 3

A 66-year-old male patient with NPC. (A) The right nasopharyngeal wall of T1WI was significantly thickened. (B) The lesion of BF was uniformly perfused, FST2WI/ASL fusion improved the lesion anatomical structure display, and the BF_mean value was 96.292 mL/100 g/min. (C) The BF_max value of the lesion was 132.17 mL/100 g/min. (D) The BF_min value of the lesion was 90.875 mL/100 g/min. (E) ADC _mean was 1.11×10⁻³ mm²/s. (F) The D_mean value was 0.891×10⁻³ mm²/s. (G) The D*_mean value was 4.46×10⁻² mm²/s. (H) Low Ki-67 expression of 10% (IHC, ×400, EnVision). NPC, nasopharyngeal carcinoma; BF, blood flow; BF_mean, mean BF; BF_max, maximum BF; BF_min, minimum BF; ADC, apparent diffusion coefficient; ADC_mean, mean ADC; D_mean, mean D; D*, pseudo-diffusion coefficient; D*_mean, mean D*.

Figure 4

A 38-year-old male patient with NPC. (A) Enhanced T1WI showed nasopharyngeal soft tissue symmetry thickening. (B) FST2WI/ASL fusion images showed obvious hyperperfusion of BF lesions, with a BF_mean value of 256.39 mL/100 g/min. (C) The BF_max value of the lesion was 358.53 mL/100 g/min. (D) BF_min of the lesion was 207.92 mL/100 g/min. (E) The D_mean value was 0.814×10⁻³ mm²/s. (F) The D*_mean value was 10.6×10⁻² mm²/s. (G) The f_mean value was 0.208. (H) The high expression of Ki-67 was 85% (IHC, ×400, EnVision). NPC, nasopharyngeal carcinoma; BF, blood flow; BF_mean, mean BF; BF_max, maximum BF; BF_min, minimum BF; ADC, apparent diffusion coefficient; D_mean, mean D; D*, pseudo diffusion coefficient; D*_mean, mean D*; f, perfusion fraction; f_mean, mean f.

Figure 5

ROC curve for BF_max in differentiating between low and high Ki-67 statuses in NPC. BF_max, maximum blood flow; ROC, receiver operating characteristic; NPC, nasopharyngeal carcinoma.