Skip to main content
NCBI home page
The British Journal of Radiology logoLink to The British Journal of Radiology
. 2022 Mar 3;95(1134):20211374. doi: 10.1259/bjr.20211374

Dynamic contrast-enhanced MRI can quantitatively identify malignant transformation of sinonasal inverted papilloma

Zheng Li 1, Mu Xian 2, Jian Guo 3, Cheng shuo Wang 4, Luo Zhang 5,✉, Junfang Xian 6,✉
PMCID: PMC10996421  PMID: 35234501

Abstract

Objective:

To investigate the diagnostic performance of quantitative and semi-quantitative parameters derived from dynamic contrast-enhanced MRI (DCE-MRI) in differentiating sinonasal inverted papilloma (SIP) from SIP with coexisting malignant transformation into squamous cell carcinoma (MT-SIP).

Methods:

This retrospective study included 122 patients with 88 SIP and 34 MT-SIP. Quantitative and semi-quantitative parameters derived from DCE-MRI were compared between SIP and MT-SIP. The multivariate logistic regression analysis was performed to identify independent indicators and construct regression model for distinguishing MT-SIP and SIP. Diagnostic performance of independent indicators and regression model were evaluated using receiver operating coefficient (ROC) analysis and compared using DeLong test.

Results:

There were significant differences in maximum slope of increase, contrast-enhancement ratio, bolus arrival time, volume of extravascular extracellular space (V e ), and rate constant (Kep) between SIP and MT-SIP (p < 0.05). There were no significant differences in initial area under the gadolinium curve (p = 0.174) and volume transfer constant (p = 0.105) between two groups. Multivariate analysis results showed that Ve and Kep were identified as the independent indicators for differentiating MT-SIP from SIP (p < 0.001). Areas under the ROC curves (AUCs) for predicting MT-SIP were 0.779 for Ve and 0.766 for Kep. The AUC of the combination of Ve and Kep was 0.831, yielding 83% specificity and 76.5% sensitivity.

Conclusion:

DCE-MRI can quantitatively differentiate between MT-SIP and SIP. The combination of Ve and Kep yielded an optimal performance for discriminating SIP from its malignant mimics.

Advances in knowledge:

DCE-MRI with quantitative and semi-quantitative parameters can provide valuable evidences for quantitatively identifying MT-SIP.

Sinonasal inverted papilloma (SIP) is a benign sinonasal neoplasm with an incidence of 0.74–1.5 per 100.000 per year. 1 However, squamous cell carcinoma can be associated with SIP in approximately 7–11% of cases either synchronously or metachronously. 2–4 SIP with coexisting malignant transformation into squamous cell carcinoma (MT-SIP) has different prognoses and treatment. Minimally invasive endoscopic surgery is the standard for treatment of SIP, while aggressive surgical resection combined with post-operative adjuvant radiotherapy or chemoradiotherapy is needed. 5–7 Therefore, it is necessary to identify MT-SIP before surgery.

Accurate diagnosis of MT-SIP remains clinically challenging pre-operatively. Biopsy is the current reference standard for quantifying the degree of cellular atypia and demonstrating MT-SIP. 1 However, biopsy is invasive with the risk of hemorrhagic complications. Moreover, misdiagnosis is frequently encountered due to sampling errors. 8,9 MR imaging including diffusion-weighted imaging (DWI) and dynamic contrast-enhanced (DCE)-MR imaging plays a vital role in distinguishing SIP from MT-SIP with specificity ranging from 78 to 93% and sensitivity ranging from 62 to 100%. 3,10 Nonetheless, the repeatability and reliability concerns have been raised in these studies because of the variation of evaluating morphological features and the type of time–intensity curve on MRI among different investigators.

Several studies have involved semi-quantitative and quantitative parameters derived from DCE-MRI to characterize the vascular properties of sinonasal tumors, yielding a high accuracy in discriminating benign from malignant sinonasal tumors. 11–13 Moreover, the quantitative nature of these imaging parameters could greatly increase reliability in the clinical setting. In addition, the pseudocolor maps of these parameters can provide visual–anatomical information on tumor boundaries, hence facilitating region of interest (ROI) selection. 14 Therefore, we hypothesized that the quantitative and semi-quantitative parameters derived from DCE-MRI can be a useful tool for the detection of MT-SIP. This study aims to investigate the diagnostic performance of quantitative and semi-quantitative parameters derived from DCE-MRI in differentiation between SIP and MT-SIP, which would provide more accurate information for identifying MT-SIP.

Methods and materials

Our institutional review board approved this retrospective case–control study. Written informed consent was waived because of the retrospective nature of the study.

SIP and MT-SIP patients

From January 2016 to December 2020, 141 consecutive patients with SIP and MT-SIP proven by pathology were retrospectively reviewed. The inclusion criteria were as follows: (1) MR scanning was performed before the biopsy and treatment. (2) DCE-MRI sequences were available. Exclusion criteria included: (1) patients with poor quality of images (n = 11) and (2) patients with the tumor less than 10 mm in short-axis diameters (n = 8). 15 Ultimately, a total of 122 patients, including 88 patients with SIP [65 male and 23 female; age range 30–80 years; median age 54 years; median tumor volume 35.5 (×103 /mm3)] and 34 patients with MT-SIP [31 male and three female; age range 31–79 years; median age 62 years; median tumor volume 32.4 (×103 /mm3)], were enrolled in this study. The specific location of tumors detected by surgery for each patient.

MRI acquisition

Sinonasal MR imaging was performed at 3.0 T MR system (68 patients with Signa HDxt and 54 patient with Discovery MR750, GE Healthcare, Milwaukee, WI) with an 8-channel head coil. The acquisition parameters of axial T 1 weighted fast spin echo imaging were repetition time (TR)/echo time (TE) = 560–590 ms/7–10 ms, slice thickness = 3.5–4 mm, slice gap = 0.3–0.4 mm, number of excitations (NEX) = 2; acquisition matrix = 320 × 256 or 288 × 224. Parameters of axial T 2 weighted fast spin echo imaging were TR/TE = 3975–4720 ms/85–89 ms, slice thickness = 3.5–4 mm, slice gap = 0.3–0.4 mm, NEX = 2; acquisition matrix = 320 × 256 or 288 × 224.

DCE-MR images were acquired via a fast-spoiled gradient recalled (FSPGR) T 1WI. Detailed imaging parameters were as follows: TR/TE = 3.9 ms/1.6 ms, slice thickness = 4–5 mm, no slice gap, flip angle = 11°, NEX = 2, acquisition matrix = 256×160. After the acquisition of one baseline phase of images, gadopentetate dimeglumine (Magnevist, Bayer Schering, Berlin, Germany) (a dose of 0.1 mmol/kg) was intravenously bolus injected via a power injector at a flow rate of 2 ml s−1 followed by a 20 ml saline flush at the same rate. Then, a total of 31 or 36 phases of images were sequentially acquired with a temporal resolution of 8 sec. After the DCE-MRI sequence, a post-contrast axial T1WI with fat saturation was obtained with parameters same as unenhanced T1WI, followed by post-contrast coronal and sagittal T 1-weighted images (TR/TE = 450–720 ms/10–12 ms; slice thickness = 3–4 mm, slice gap = 5.5 mm; NEX = 2; acquisition matrix = 320 × 256).

DCE-MRI derived parameters

DCE-MRI data were postprocessed using Advantage Workstation (v. 4.6, GE Healthcare, Milwaukee, WI). Signal intensity changes in internal carotid were automatically measured for arterial input function (AIF). Then, pixel-based maps for DCE parameters were calculated using GenIQ software (GE Healthcare, Milwaukee, WI).

All quantitative analysis on DCE-MR images were based on two-compartment modified Tofts model. The calculated quantitative parameters were the transfer constant between intra- and extravascular extracellular spaces (Ktrans; min−1), the volume of the extravascular extracellular space (Ve), and the transfer constant from the extravascular extracellular space to plasma (Kep; min−1, Kep = Ktrans/ Ve).

The average signal intensity within the pixel was plotted against time, and TICs were constructed. The semi-quantitative parameters were calculated on a pixel-by-pixel basis from the TICs, including maximum slope of the steepest part of the concentration curve (MaxSlope; mmol/s); the contrast-enhancement ratio (CER) = (peak signal-baseline signal)/(baseline signal); the bolus arrival time (BAT; s)=the time from the start of contrast injection to tracer bolus arrival at a lesion; and the initial area under the signal intensity–time curve (IAUGC) = (area under the tissue concentration curve from BAT to 60 s from the start of contrast injection)/(area under the AIF concentration curve from BAT to 60 s from the start of contrast injection).

Imaging analysis

Two readers (Reader 1 and Reader 2 with 8 and 20 years of experience in head and neck radiology, respectively) were blinded to the pathological results and delineated two-dimensional ROIs manually. The ROIs covered the largest single-slice at the level of maximum diameter of tumor based on DCE-MR images, avoiding the obvious necrotic, cystic, and hemorrhagic areas inside tumor and surrounding blood vessels (Figure 1). Edges of lesions were also excluded from ROI to minimize the effect of partial volume averaging. The measurements of two readers were used to assess the interobserver consistency. For intraobserver consistency, Reader 1 repeated the process of ROI placement with a month interval, and the average results of Reader 1 were used for statistical analysis.

Figure 1.

Figure 1.

Open in a new tab

Illustrations of ROI placement. A 2D-ROI covered a slice with the largest diameter of tumor based on DCE sequence, avoiding the obvious necrotic, cystic, and hemorrhagic areas inside tumor and surrounding blood vessels. 2D, two-dimensional; DCE, dynamic contrast-enhanced; ROI, region of interest.

Statistical analyses

Inter- and intraobserver consistency of each quantitative parameters based on ROIs were assessed by using the intraclass correlation coefficient (ICC). The consistency was defined as follows: <0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and >0.80, excellent.

The Kolmogorov–Smirnov test was used to test the normality for each continuous variable, after which the variable was expressed as median (interquartile range) when data non-normally distributed. Mann–Whitney U test (non-normally distributed) was used to compare parametric values between SIP and MT-SIP groups. The significantly different parameters were involved in a forward stepwise multivariable logistic regression analysis for determining independent diagnostic parameters and the best logistic regression model. Receiver operating characteristic (ROC) curve analyses were performed to evaluate the diagnostic performance of independent diagnostic parameters and regression model in differentiating MT-SIP from SIP. The optimal threshold of each independent diagnostic parameter was determined by the maximum Youden index (Youden index = sensitivity + specificity −1). Differences of area under the ROC curve (AUC) were compared using the DeLong’s test.

All statistical analyses were performed with statistical software (SPSS v. 19.0, SPSS, Chicago, IL; and Medcalc v. 9.0, MedCalc Software, Mariakerke, Belgium). The p < 0.05 values were considered significant.

Results

Tumor location

SIPs were found in the maxillary sinus in 31 cases (35.2%), in the nasal cavity in 28 cases (31.8%), in the ethmoid sinus in 15 cases (17.1%), in the frontal sinus in 9 cases (10.2%), and in the sphenoid sinus in 5 cases (5.7%). MT-SIPs were identified in the maxillary sinus in 16 cases (47.1%), in the nasal cavity in 9 cases (26.5%), in the ethmoid sinus in 6 cases (17.6%), and in the sphenoid sinus in 3 cases (8.8%).

Comparison of quantitative and semi-quantitative parameters derived from DCE-MRI between SIP and MT-SIP

Maxslope, CER, BAT and Ve were significantly higher in SIP than those in MT-SIP (p < 0.05). Kep was significantly lower in SIP than that in MT-SIP (p < 0.05) (Table 1). There was no significant difference in IAUGC and Ktrans (p > 0.05) between two groups (Table 1).

Table 1.

Comparisons of DCE-MRI parameters between SIP and MT-SIP and ICCs of intra- and interobserver consistency

Parameters SIP (n = 88) MT-SIP (n = 34) p-value Interobserver consistency Intraobserver consistency
  ICC 95% CI ICC 95% CI
Maxslope(mmol/s) 0.03 (0.02,0.05) 0.02 (0.02,0.04) 0.003 0.614 0.490–0.713 0.762 0.660–0.834
CER 1.76 (1.59,2.11) 1.50 (1.22,1.79) <0.001 0.816 0.756–0.868 0.903 0.864–0.931
IAUGC 0.30 (0.21,0.43) 0.31 (0.25,0.63) 0.174 0.863 0.810–0.902 0.938 0.913–0.956
BAT(s) 33.09 (23.25,42.43) 24.08 (19.44,33.64) 0.022 0.975 0.965–0.983 0.981 0.973–0.987
Ktrans(min−1) 0.37 (0.24,0.54) 0.44 (0.29,0.66) 0.105 0.954 0.935–0.968 0.939 0.914–0.957
Ve 0.66 (0.56,0.84) 0.49 (0.36,0.58) <0.001 0.843 0.770–0.881 0.850 0.792–0.893
Kep(min−1) 0.62 (0.45,0.88) 1.05 (1.33,0.79) <0.001 0.963 0.947–0.974 0.982 0.975–0.988
Open in a new tab

BAT, the bolus arrival time; CER, the contrast-enhancement ratio; CI, confidence interval; IAUGC, the initial area under the signal intensity–time curve; ICC, intraclass correlation coefficient; MT-SIP, malignant transform of sinonasal inverted papillomas; ROI, region of interest; SIP, sinonasal inverted papillomas.

Inter- and intraobserver agreement

The inter- and intraobserver agreement for each parameter derived from DCE-MRI were indicated in Table 1. The inter- and intraobserver agreement were both excellent (ICCs = 0.816–0.982) for CER, BAT, IAUGC, Ktrans, Ve, and Kep. Good inter- and intraobserver agreement were shown in Maxslope (ICCs = 0.614, 0.762).

Multivariate logistic regression analysis

Multivariate logistic regression analysis demonstrated that Ve [Odds ratio (OR)=0.004; 95% confidence interval (95% CI)=0.001–0.065; p < 0.001] and Kep (OR = 3.852; 95%CI = 1.609–9.219; p = 0.002) were independent indicators for differentiating MT-SIP from SIP (Figures 2 and 3, Supplementary Material 1).

Figure 2.

Figure 2.

Open in a new tab

A 43-year-old male with malignant inverted papilloma transformation in the right maxillary sinus. Pseudocolor map of Ve(a) shows a mean value of Ve of malignant inverted papilloma transformation (red arrow) is 0.48, lower than the cutoff value of 0.58. Pseudocolor map of Kep (b) shows a mean value of Kep of malignant inverted papilloma transformation (red arrow) is 1.03(min−1), higher than the cut-off value of 0.75 (min−1).

Figure 3.

Figure 3.

Open in a new tab

A 60-year-old male with inverted papilloma in the left maxillary sinus. Pseudocolor map of Ve(a) shows a mean value of Ve of inverted papilloma (red arrow) is 0.66, higher than the cut-off value of 0.58. Pseudocolor map of Kep (b) shows a mean value of Kep of inverted papilloma (red arrow) is 0.57 (min−1), lower than the cut-off value of 0.75 (min−1).

ROC curve analysis

The AUC, sensitivity, specificity, positive-predicted value, and negative-predicted value of each independently diagnostic parameter and regression model discriminating SIP and MT-SIP were showed in Table 2 and Figure 4. Using a Ve value of 0.582 or a Kep value of 0.747 as a threshold value for diagnosis, the sensitivity and specificity were 79.4 and 68.2%, respectively. The AUC for diagnosis of Ve and Kep were 0.779 (95%CI = 0.695–0.849) and 0.766 (95%CI = 0.681–0.838), respectively.

Table 2.

ROC analysis for Ve, Kep, and their combination model in distinguishing between SIP and MT-SIP

Parameters Cut-off SEN (%) SPE (%) PPV (%) NPV (%) AUC (95% CI)
Ve 0.582 79.4 68.2 49.1 89.6 0.779 (0.695–0.849)
Kep (min−1) 0.747 79.4 68.2 49.1 89.6 0.766 (0.681–0.838)
Ve + Kep 76.5 83.0 63.4 90.1 0.831 (0.752–0.892)
Open in a new tab

AUC, area under the curve; IAUGC, the initial area under the signal intensity–time curve; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristics; ROI, region of interest; SEN, sensitivity; SPE, specificity.

Figure 4.

Figure 4.

Open in a new tab

ROC curves and AUC (in brackets) for Ve, Kep, and their combination model.AUC, area under the curve; ROC, receiver operating coefficient; ROI, region of interest.

When combined Ve and Kep, the specificity increased to 83% from 68.2%, while the sensitivity decreased to 76.5% from 79.4%. The combination of Ve and Kep showed the largest AUC of 0.831 (95%CI = 0.752–0.892) in discrimination of SIP and MT-SIP. However, there was no significant difference of AUC among them (Ve vs Kep, p = 0.846; combined model vs Ve, p = 0.0654; combined model vs Kep, p = 0.1274). The Hosmer-Lemeshow goodness-of-fit test result was not significantly different for the combined model.

Discussion

Given its aggressive nature and poor prognosis, MT-SIPs require radical treatment strategies. Hence, it is of great importance to discriminate MT-SIP from SIP before making a treatment decision. This retrospective study compared a variety of parameters derived from DCE-MRI and suggested that Maxslope, CER, BAT, Ve and Kep were significantly different between SIP and MT-SIP, with excellent to good repeatability. Moreover, Ve and Kep were independent indicators for discriminating MT-SIP from SIP, and combination of Ve and Kep yielded an optimal diagnostic performance with an AUC of 0.831. Hence, DCE-MRI can be a quantitative technique for detecting MT-SIP non-invasively.

DCE-MRI was suggested to reflect microvascular environment in sinonasal cancers using pharmacokinetic models with quantitative parameters. 11,16 However, there was a paucity of quantitative parameters derived from DCE-MRI that focused on differentiation between SIP and MT-SIP. Kep is a quantitative parameter that refers to the contrast agent transfer from the tissue space to the blood and has been shown to be positively associated with vascular permeability of tumor. 17 Our study showed that Kep values were significantly larger in MT-SIP than that in SIP, suggesting that MT-SIP has incomplete microvascular circulation and high permeability. However, Ktrans, as another quantitative parameter relates to vascular permeability, showed no significant difference between MT-SIP and SIP. One possible reason for this phenomenon may be due to the fact that Ktrans can be affected by microvascular density of tumor beside the vascular permeability. 18 For example, a high microvascular density in the peripheral portion of SIP would definitely elevate its mean value of Ktrans, and therefore the Ktrans of SIP may be comparable with that of MT-SIP, which mainly attributed to high permeability. 18

While noting the limitations posed by lacking reproducibility, previous studies concluded that DWI with apparent diffusion coefficient value can reflect cellularity of tumors and can be a highly sensitive tool for differentiation between SIP and MT-SIP. 3 However, it has not developed into a standalone diagnostic tool due to its relatively low specificity. Moreover, Ve is a quantitative parameter derived from DCE-MRI, which can simultaneously quantify cellularity and vascular properties of tumor. 19 Depending mainly on the cell density and the composition of the extravascular extracellular matrix, Ve value is correlated inversely with the tumor cellularity and positively with the volume of extravascular extracellular space. 20 Our study showed that Ve was significantly lower in MT-SIP than that in SIP. This result was very similar to that of the previous study 3 and suggested that MT-SIP has a higher cellularity and a lower extracellular space compared to SIP.

Furthermore, this study demonstrated that Kep and Ve can be independently used to distinguish SIP from MT-SIP. The change in vascular permeability and cellularity seemed to be the major factors in malignant transformation of SIP. This result contained an important message that monitoring malignant transformation of SIP might be objectively estimated by quantitative parameters derived from DCE-MRI. In addition, ROC analysis suggested a powerful prediction of Kep and Ve with an AUC of 0.766 and 0.779, respectively. However, a combination of these two parameters was needed for improving the specificity in predicting MT-SIP.

Combination of Ve and Kep showed a satisfactory performance in predicting MT-SIP with an AUC of 0.831. Further, the combined model increased the specificity of DCE-MRI, although the sensitivity slightly decreased. These results indicated that comprehensive evaluation of quantitative parameters may further refine our ability to differentiate benign from malignant sinonasal tumors on DCE-MRI and may reduce the need for biopsy.

Semi-quantitative parameters derived from DCE-MRI may help for reflecting divergent hemodynamic patterns of tumors. 21 Further, the semi-quantitative data are unlikely to be influenced by variation of scanning protocols and is easier in post-processing compared with the quantitative data. 22 In our study, Maxslope and CER in SIP significantly were higher than those in MT-SIP, suggesting that degree of enhancement in SIP was higher than MT-SIP. While, shorter BAT for MT-SIP might associate with shunt formation and low-resistance and high-flow pathways within tumor. However, none of semi-quantitative parameters were independently associated with MT-SIP based on multivariate logistic regression analysis. This finding, to some degree, suggested that the diagnostic performance of semi-quantitative parameters was inferior to that of quantitative parameters in detecting MT-SIP.

Limitations

Our study had several limitations. First, the study included a relatively small population from a single institution. Multicentric studies with larger patient populations were required to further validate our results. Second, we did not report on the conventional MRI features for the differentiation of tumors, the assessment of which was not within the scope of our study. Finally, obtaining the completed surgical specimen for histopathological analysis is difficult in clinical practice, thus a head-to-head evaluation of images and histopathological findings of areas covered by ROI had not been performed.

Conclusions

In conclusion, DCE-MRI can quantitatively differentiate between MT-SIP and SIP. Ve and Kep were independent indicators for discriminating MT-SIP from SIP with high repeatability. The combination of Ve and Kep may obtain an optimal performance for discriminating SIP from its malignant mimics.

Supplementary Material

bjr.20211374.suppl-01
bjr.20211374.suppl-01.docx (1.7MB, docx)

Footnotes

Acknowledgments: The authors thank Kai Cao (Beijing Tongren Hospital) who helped with the statistics.

Contributor Information

Zheng Li, Email: li490354899@126.com, Department of Radiology, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

Mu Xian, Email: mu_xian@126.com, Department of Otolaryngology Head and Neck Surgery, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

Jian Guo, Email: jiangjoojle@163.com, Department of Radiology, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

Cheng shuo Wang, Email: wanges830@126.com, Department of Otolaryngology Head and Neck Surgery, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

Luo Zhang, Email: dr.luozhang@139.com, Department of Otolaryngology Head and Neck Surgery, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

Junfang Xian, Email: cjr.xianjunfang@vip.163.com, Department of Radiology, Beijing Tongren Hospital, Capital Medical University, Beijing, China .

REFERENCES

  • 1. Lisan Q, Laccourreye O, Bonfils P . Sinonasal inverted papilloma: from diagnosis to treatment . Eur Ann Otorhinolaryngol Head Neck Dis 2016. ; 133: 337 – 41 : S1879-7296(16)30054-0 . doi: 10.1016/j.anorl.2016.03.006 [DOI] [PubMed] [Google Scholar]
  • 2. Lee JJ, Roland LT, Licata JJ, Orlowski HLP, Jiramongkolchai P, et al. . Morphologic, intraoperative, and histologic risk factors for sinonasal inverted papilloma recurrence . Laryngoscope 2020. ; 130: 590 – 96 . doi: 10.1002/lary.28078 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Yan CH, Tong CCL, Penta M, Patel VS, Palmer JN, et al. . Imaging predictors for malignant transformation of inverted papilloma . Laryngoscope 2019. ; 129: 777 – 82 . doi: 10.1002/lary.27582 [DOI] [PubMed] [Google Scholar]
  • 4. Udager AM, McHugh JB, Goudsmit CM, Weigelin HC, Lim MS, et al. . Human papillomavirus (HPV) and somatic EGFR mutations are essential, mutually exclusive oncogenic mechanisms for inverted sinonasal papillomas and associated sinonasal squamous cell carcinomas . Ann Oncol 2018. ; 29: 466 – 71 : S0923-7534(19)35054-9 . doi: 10.1093/annonc/mdx736 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Re M, Gioacchini FM, Bajraktari A, Tomasetti M, Kaleci S, et al. . Malignant transformation of sinonasal inverted papilloma and related genetic alterations: a systematic review . Eur Arch Otorhinolaryngol 2017. ; 274: 2991 – 3000 . doi: 10.1007/s00405-017-4571-2 [DOI] [PubMed] [Google Scholar]
  • 6. Contrera KJ, Woody NM, Rahman M, Sindwani R, Burkey BB, et al. . Clinical management of emerging sinonasal malignancies . Head Neck 2020. ; 42: 2202 – 12 . doi: 10.1002/hed.26150 [DOI] [PubMed] [Google Scholar]
  • 7. Lee JJ, Peterson AM, Embry TW, Wamkpah NS, Kallogjeri D, et al. . Survival outcomes of de novo vs inverted papilloma-associated sinonasal squamous cell carcinoma: a systematic review and meta-analysis . JAMA Otolaryngol Head Neck Surg 2021. ; 147: 350 – 59 . doi: 10.1001/jamaoto.2020.5261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Marri UK, Das P, Kalaivani M, Srivastava DN, et al. . Noninvasive staging of liver fibrosis using 5-minute delayed dual-energy CT: comparison with US elastography and correlation with histologic findings . Radiology 2021. ; 298: 600 – 608 . doi: 10.1148/radiol.2021202232 [DOI] [PubMed] [Google Scholar]
  • 9. Schreiber A, Rampinelli V, Ferrari M, Mattavelli D, Farina D, et al. . Diagnostic reliability of pretreatment biopsy in malignant nasoethmoidal tumors: A retrospective study of 77 cases . Laryngoscope 2018. ; 128: 1772 – 77 . doi: 10.1002/lary.27077 [DOI] [PubMed] [Google Scholar]
  • 10. Zhang L, Fang G, Yu W, Yang B, Wang C, et al. . Prediction of malignant sinonasal inverted papilloma transformation by preoperative computed tomography and magnetic resonance imaging . Rhinology 2020. ; 58: 248 – 56 . doi: 10.4193/Rhin19.240 [DOI] [PubMed] [Google Scholar]
  • 11. Xian J, Du H, Wang X, Yan F, Zhang Z, et al. . Feasibility and value of quantitative dynamic contrast enhancement MR imaging in the evaluation of sinonasal tumors . Chin Med J (Engl) 2014. ; 127: 2259 – 64 . [PubMed] [Google Scholar]
  • 12. Wang X, Zhang Z, Chen X, Li J, Xian J, et al. . Value of magnetic resonance imaging including dynamic contrast-enhanced magnetic resonance imaging in differentiation between inverted papilloma and malignant tumors in the nasal cavity . Chin Med J (Engl) 2014. ; 127: 1696 – 1701 . [PubMed] [Google Scholar]
  • 13. Wang X-Y, Yan F, Hao H, Wu J-X, Chen Q-H, et al. . Improved performance in differentiating benign from malignant sinonasal tumors using diffusion-weighted combined with dynamic contrast-enhanced magnetic resonance imaging . Chin Med J (Engl) 2015. ; 128: 586 – 92 . doi: 10.4103/0366-6999.151649 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Cheng J, Shao S, Chen W, Zheng N, et al. . Application of diffusion kurtosis imaging and dynamic contrast-enhanced magnetic resonance imaging in differentiating benign and malignant head and neck lesions . J Magn Reson Imaging 2022. ; 55: 414 – 23 . doi: 10.1002/jmri.27885 [DOI] [PubMed] [Google Scholar]
  • 15. Xiao Z, Zhong Y, Tang Z, Qiang J, Qian W, et al. . Standard diffusion-weighted, diffusion kurtosis and intravoxel incoherent motion MR imaging of sinonasal malignancies: correlations with ki-67 proliferation status . Eur Radiol 2018. ; 28: 2923 – 33 . doi: 10.1007/s00330-017-5286-x [DOI] [PubMed] [Google Scholar]
  • 16. Song C, Cheng P, Cheng J, Zhang Y, Sun M, et al. . Differential diagnosis of nasopharyngeal carcinoma and nasopharyngeal lymphoma based on DCE-MRI and RESOLVE-DWI . Eur Radiol 2020. ; 30: 110 – 18 . doi: 10.1007/s00330-019-06343-0 [DOI] [PubMed] [Google Scholar]
  • 17. Cheng Z, Wu Z, Shi G, Yi Z, Xie M, et al. . Discrimination between benign and malignant breast lesions using volumetric quantitative dynamic contrast-enhanced MR imaging . Eur Radiol 2018. ; 28: 982 – 91 . doi: 10.1007/s00330-017-5050-2 [DOI] [PubMed] [Google Scholar]
  • 18. Qian W, Xu X-Q, Hu H, Su G-Y, Wu J-F, et al. . Dynamic contrast-enhanced MRI in orbital lymphoproliferative disorders: effects of region of interest selection methods on time efficiency, measurement reproducibility, and diagnostic ability . J Magn Reson Imaging 2018. ; 47: 1298 – 1305 . doi: 10.1002/jmri.25859 [DOI] [PubMed] [Google Scholar]
  • 19. Dong Ji X, Yan S, Xia S, Guo Y, Shen W, et al. . Quantitative parameters correlated well with differentiation of squamous cell carcinoma at head and neck: a study of dynamic contrast-enhanced MRI . Acta Radiol 2019. ; 60: 962 – 68 . doi: 10.1177/0284185118809543 [DOI] [PubMed] [Google Scholar]
  • 20. Meyer H-J, Hamerla G, Leifels L, Höhn AK, Surov A, et al. . Histogram analysis parameters derived from DCE-MRI in head and neck squamous cell cancer - associations with microvessel density . Eur J Radiol 2019. ; 120: 108669 : S0720-048X(19)30319-5 . doi: 10.1016/j.ejrad.2019.108669 [DOI] [PubMed] [Google Scholar]
  • 21. Ota Y, Liao E, Capizzano AA, Kurokawa R, Bapuraj JR, et al. . Diagnostic role of diffusion-weighted and dynamic contrast-enhanced perfusion MR imaging in paragangliomas and schwannomas in the head and neck . AJNR Am J Neuroradiol 2021. ; 42: 1839 – 46 . doi: 10.3174/ajnr.A7266 [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  • 22. Lee JY, Cheng KL, Lee JH, Choi YJ, Kim HW, et al. . Detection of local recurrence in patients with head and neck squamous cell carcinoma using voxel-based color maps of initial and final area under the curve values derived from DCE-MRI . AJNR Am J Neuroradiol 2019. ; 40: 1392 – 1401 . doi: 10.3174/ajnr.A6130 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

bjr.20211374.suppl-01
bjr.20211374.suppl-01.docx (1.7MB, docx)

Articles from The British Journal of Radiology are provided here courtesy of Oxford University Press

RESOURCES