Does Diffusion Magnetic Resonance Imaging (DWI) has Role in Irradiated Laryngeal Carcinoma?

Abstract

DWI involves acquisition of signal of movement of water proton in cellular spaces of body (Brownian motion). It includes qualitative method either restricted or facilitated and quantitive method which is apparent diffusion coefficient value(ADC) which is related to proportion of extracellular and intracellular components of the tissue., ADC is calculated with use of at least two b value more accurate using more DWI with different b value,ADC levels is low in increased tissue cellularity, as malignancy., ADC levels is high in non-tumoral tissue alterations such as direct endoscopy oedema, radiotherapy necrosis are expected to have minimal cellularity. ADC is most accurate in the detection of malignancy versus tissue edema or radionecrosis the aim of study to assess value of ADC as regarding measuring sensitivity and specificity and accuracy to differentiate tumor recurrence from radionecrosis. This study includes 36 patients who were suspected patients of tumor recurrence after radiotherapy; it is a prospective randomized comparative clinical trial. The patients were assessed using direct laryngoscopic examination under general anaesthesia and biopsy, and diffusion weighted image on the neck (b0 and b1000), ADC map and ADC value measured al lesion and normal tissues and compared with pathology results. ADC value (mean 0.93 ± 0.30 X 10^–3 mm²/s) in patients had recurrent carcinoma was significantly lower (P < .0001) than the mean ADC of normal tissue in the same patients (1.26 ± 0.134) while mean ADC of tumour recurrence (P < .0001) was lower than mean ADC value of radio necrosis (1.63 ± 0.21 × 10^–3 mm²/s). MRI ADC value is a sensitive and non-invasive method in detection of a recurrent laryngeal lesion from radionecrosis.

Introduction

Tumors can be detected using traditional imaging techniques by looking for anatomic distortion or changes in tissue appearance. Post-treatment sequel or radionecrosis or infection may resemble malignancy [1]. The DWI technique gives information on movements of water molecules in tissues [2]. Due to restrictions, such as those imposed by cell membranes, fibers and macromolecules, completely unrestricted diffusion of molecules does not exist in biological tissues [3]. In the human body, the apparent diffusion coefficient relates to the degree of translational diffusion of molecules detected (ADC). ADCs are expected to differ depending on tissue microstructures and pathophysiologic conditions [4], ADC is usually expressed (× 10^–3) as square millimetres per second, ADC calculation is an automated process calculated with at least two b value, better with more b values, the final image with different ADC value calculated for each pixel of an image. DWI has been successfully used in brain lesions assessment but uses in in laryngeal lesions have been limited due to complexity of anatomy of this area as containing adjacent air and bone, motion artifacts from the airway and swallowing and dental prosthesis, [5]. However, novel applications are being developed that allow DWI to be used in laryngeal carcinoma for delineate recurrence or residual tumour, and post treatment sequel [2, 6, 7]. In malignant tumors,ADC levels will be low due to tissue hyper cellularity The ADC level is high in tissue expected to have low cellularity as in post direct endoscopy oedema and radiotherapy swelling or necrosis. [8]. DWI used to assess the larynx for tumour recurrence or residual carcinoma or post radiotherapy oedema [1, 9–11].

This study aims to assess role of DWI generated from diffusion MRI in detecting recurrent/residual laryngeal carcinoma and post radiotherapy oedema.

Materials and Methods

This study was conducted in the otorhinolaryngology department and radio diagnosis department in Assiut University Hospital From April 2019 to April 2021, The study is prospective randomized comparative clinical trial. The study was done after obtaining approval from the Medical Ethics Committee, Faculty of Medicine, Assiut University. All cases who participated in the study gave verbal consent for DWI and written consent for endoscopy under general anaesthesia, and all data was kept confidential.

Inclusion criteria

The current study enclosed 36 patients coming for follow-up of recurrent/residual laryngeal malignant neoplastic disease as detected by telescopic examination using 90 0 telescopic examination. The study includes irradiated patients (the total radiation dose ranged from 60 to 70 Gy given by conventional fractionation (1.8–2.0 Gy per fraction), with 5 fractions per week over 6–7.5 weeks. The filed arrangement varied according to the site of the primary lesions. The size of the irradiated fields was chosen according to the extension of the tumor, stage, and probability of lymphatic spread. MR assessments of those patients were 3–6 months once the completion of their treatment program. Patients were subjected to history taking to sight new-onset laryngeal symptoms like dysphonia, stridor dysphagia, or worsening of symptoms once chemo/radiotherapy in patients antecedently diagnosed as recurrent laryngeal carcinoma. Complete otorhinolaryngology examination was performed for each patient, as well as a telescopic examination of the speech organ to verify any abnormality as lots or any modification of the previous examination of patients (90-degree Storz ® endoscope (Germany) 8 × 179 mm autoclavable stainless steel easy for maintenance).

Exclusion criteria

Any patient with history of pacemakers, cochlear implants excluded from the study.

Irradiated patient after total laryngectomy.

MRI Protocol

Conventional Protocol

Patients were subjected to. MRI with diffusion on the neck with emphasis on the larynx was done. In our study, an MRI examination was performed by (1.5 Tesla) superconducting MR imager (Achieva, Philips Medical Systems, The Netherlands B.V.). The study was performed with the patient lying on his back and thus the patient's airway parallels to the table. All patients underwent DW-MRI which was obtained in axial plane and performed without injection of contrast material, To acquire an optimal MRI examination adequate psychological preparation of the patient, confirmation for the absence of any paramagnetic material perform. A circularly polarized surface coil was placed over the neck used, The receiver coil should be separated from the chest wall to prevent movement; however, it must be near the neck. Recommended protocols include:

1. Sagittal T2WI images are obtained initially to prescribe the location of transverse images.

2. 2-transverse T1 –weighted turbo spin-echo (TSE) transverse images extended from hyoid bone (approximately the level of the third cervical vertebra) to the lung apex. {a repetition time (TR) of 430 ms and echo time (TE) of 12 ms images}, axial T2-weighted sequences, T2 turbo spin-echo (TSE) a repetition time (TR) of 4252 ms and echo time (TE) of 100 ms} sequences with or without spectral presaturation with inversion recovery (SPIR) with a slice thickness of 4 mm with a 0.5 mm intersection gap, field of view(FOV) 23 × 25 and matrix size 228 × 512

3. Coronal T2 & Short-TI Inversion Recovery (STIR) images were obtained for the evaluation of masses that extended below the cervicothoracic Junction.

C- Diffusion MRI: DWI were acquired single-shot echo-planar image (EPI) with 20 slices in the transverse plane, 5 mm slice thickness, 0.5 mm intersection gap, FOV 23 × 28 cm, {a repetition time (TR) of 4000–4280 ms and echo time (TE) of 94–110 ms. and matrix size 94 × 160 The images were acquired using b-values (b 0 and 1000 mm2/s.). Apparent diffusion coefficient (ADC) maps were automatically calculated by MRI machine software and included in the sequence (ADC maps were measured from many images taken by different b values. ADCis the sum of motion in three-dimensional planes (three orthogonal directions), ADCwas measured by applying multiple regions of interest (ROIs) over normal laryngeal tissue excluding any suspected lesions, also ROI was applied over lesions and ADCwas measured. ROI applied by radiologist and median size of 30 mm2 to calculate ADCvalue on ADCmaps. In addition to conventional techniques, DW- MRI and ADC maps were obtained, The hypothesis of this study was that adding DW- MRI and ADC map to the routine MRI protocol used in patients with clinically suspected, laryngeal mass by telescopic examination, will result in an increase in its diagnostic performance, and to prove this hypothesis we compared the results of conventional MRI and DW MRI.

Endoscopy and Biopsy

Direct laryngoscopic examination of the larynx under general anesthesia with biopsy taking from suspicious lesions Laryngoscopic examination under general anesthesia using orotracheal tube or through tracheostomy tube.Rigid endoscope used were. Histopathological examination of biopsies was done with a correlation between radiological results of DW-MRI and pathological results.

Statistical Analysis

In this study we use IBM compatible personal computer with SPSS statistical package for social science version 23 (IBM Corp., Armonk, NY, US). Descriptive statistics included qualitative data that were expressed as number (No.) and percentage (%). Quantitative data were expressed as mean (X), standard deviation (SD), median, and range. Paired t-test to evaluate the mean ADC at radio necrosis, tumor recurrence, and normal laryngeal tissues in the same patients. Independent t-test to evaluate the mean ADC at tumor recurrence and radionecrosis. The probability (p value) was considered significant when (P < 0.5).

P ≤ 0.05 was thought to be significant. P value > 0.0001 is highly significant; we had used the receiver operative characteristic (ROC) curve to calculate sensitivity, specificity; Positive predictive value (PPV), Negative predictive value (NPV) and accuracy all were calculated to calculate cut off value of ADC.

Results

Thirty six patients were included in this study suspected of tumor recurrence or coming for follow up after 3–6 month of radiotherapy for laryngeal carcinoma. Their ages ranged from 50 to 85 years with a mean age of 62.33 years SD ± 9.47.

Histopathological examination revealed 24 patients with tumor recurrence and twelve patients with radionerosis.

Regarding the site of involvement by direct laryngoscopy under general anesthesia, we found that transglottic lesion found in 18 patients (44.44%) and glottic lesion found in 12 cases (33.33%) while subglottic lesion found in four cases (11.1%).

Conventional MRI (T2 weighted image) was stated that hyperintense lesion 25 patients (63.4%) and isointese lesions 11 patients, while diffusion MRI (qualitative method)) was stated that restricted diffusion in 25 (69.4. %), ADC value (1.27 ± 0.131) with range (1.1–1.5) (Table 1).

Table 1.

Data of the patients by conventional MRI and DW-MRI

Diffusion	No. (n = 36)	%
Facilitated	11	30.6
Restricted	25	69.4
Signal intensity on T2
Hyperintense	25	69.4
Iso intense	11	30.6
ADC value
Range	1.1	1.5
Mean ± SD	1.27 ± 0.131

ADC (mean ± SD), Signal intensity on T2 Diffusion represented by the number of patients, percentage %

ADC value (mean 0.93 ± 0.30 X 10^–3 mm²/s) in patients had recurrent carcinoma was significantly lower (P < 0.0001) than the mean ADC of normal tissue in the same patients (1.26 ± 0.134) while mean ADC of tumour recurrence (P < 0.0001) is lower than mean ADC value of radio necrosis (1.63 ± 0.21 × 10^–3 mm²/s) (Table 2).

Table 2.

mean ADC value in recurrent, radio necrosis, and normal tissue

Lesion ADC Mean ± SD	Normal ADC Mean ± SD	Paired t test	P value
Recurrence (24) 0.93 ± 0.30	1.26 ± 0.134	4.660	< 0.001**
Necrosis (12) 1.63 ± 0.21	1.28 ± 0.127	4.329	0.001**

Independent t test of ADC of tumor recurrence of radionrosis = 7.190 P value < 0.0001

ADC mean ± SD, P value. paired t test, Independent t test

ADC value (mean 0.93 ± 0.30 X 10–³ mm²/s) in patients had recurrent carcinoma was significantly lower (P < .0001) than the mean ADC of normal tissue in the same patients (1.26 ± 0.134) while mean ADC of tumour recurrence (P < .0001) was lower than mean ADC value of radio necrosis (1.63 ± 0.21 × 10^–3 mm2/s)

The cut off value of the ADCusing receiver operating characteristics ≥ 1.2X 10^−.3 mm²/s with the area under the curve 0.976. with highly significant P value > 0.0001, with sensitivity 95.83% and specificity 91.67 and accuracy of 93.75%. (Table 3, Fig. 1), while, diffusion (restricted versus facilitated technique) with sensitivity 91.67% and specificity 75% and accuracy 83.3% with significant p value 0.001 with the area under the curve 0.833 while conventional MRI show sensitivity of 79.17% and specificity of 50% and accuracy of 64.6% with an insignificant p-value with the area under the curve 0.646 (Table 3).

Table 3.

Comparison between pathology result and DW-MRI

	AUC	Sensitivity	Specificity	PPV	NPV	Accuracy	P value
ADC value	0.976	95.83	91.67	95.8	91.7	93.75	< 0.001**
Diffusion	0.833	91.67	75	88	81.8	83.3	< 0.001**
Signal intensity on T2	0.646	79.17	50	76	54.5	64.6	0.092

Fig. 1.

The ROC curve for use of ADC values in differentiating tumor recurrence from radionecrosis is (cut off value is ≥ 1.2 the area under the ROC curve 0.976, 95% CI: 0.888–1.035

Comparison between conventional MRI (signal intensity on T2) and diffusion MRI (both qualitative and quantitive methods (ADC value) in comparison with pathology result, we calculate sensitivity and specificity, positive predictive value negative predictive value, and accuracy on Roc curve (Table 3) and (Fig. 1).

Discussion

The DWI is a novel functional approach method that now has a role in laryngeal cancer imaging, provides information about tissue cellularity and cell membrane integrity [1]. DWI discriminates recurrent /residual versus postreatment sequel, low ADC values seen within malignant tumours due to hypercellularity. The ADC level is high in tissue expected to have low cellularity as post direct endoscopy edema, radiotherapy swelling or necrosis. [7]. DWI also appears to be more effective than PET (Positron emission tomography) at distinguishing between neoplastic illness and inflammatory alterations. DWI had lower cost than PET, which need tracing material, and it is ionizing radiation [8].

DW-MRI has become useful tool for discriminating tumours; DW-MRI involves the acquiring of magnetic resonance signals associated with the diffusion of water nucleon in tissue (Brownian motion, this motion is basically random and reduced by a structural barrier like cell membranes, fibers and macromolecule. The signal obtained with (DWI) is net displacement of water molecule e.g. the extracellular, intracellular, and intravascular areas, restricted diffusion in lived tissue is directly proportional to tissue physiological state and integrity of cell membranes and it may be quantified with ADC map [12].

Calculation of ADC (Quantitiuve assessment) measured by DW-MRI signal in extracellular and intracellular components of tissue and,.Koh and Collins 2007 quoted that there is an inverse relationship between tumour and ADC in in malignancy [13].

ADC value (mean 0.93 ± 0.30 X 10^–3 mm²/s) in patients had recurrent carcinoma was significantly lower (P < 0.0001) than the mean ADC of normal tissue in the same patients (1.26 ± 0.134) while mean ADC of tumour recurrence (P < 0.0001) was lower than mean ADC value of radio necrosis (1.63 ± 0.21 × 10^–3 mm2/s) (Table 2) also we found that cut off value of the ADCusing receiver operating characteristics ≥ 1.2X 10^−.3 mm2/s with the area under the curve 0.976. with highly significant P value > 0.0001, with sensitivity 95.83% and specificity 91.67 and accuracy of 93.75%. (Table 3, Fig. 1), while conventional MRI show sensitivity of 79.17% and specificity of 50% and accuracy of 64.6% with an insignificant p-value with the area under the curve 0.646. (Table 3).

The microarchitecture of tissue either facilitated or restricted, Many causes of facilitated diffusion as intratumoral swelling and cystic ( necrosis or haemorrhage) of tumor,, restricted diffusion is in cellular portion of the neoplasm [14]. As regards qualitative assessment, visual assessment of DW-MRI was applied for neoplasm tissue detection and shape, direct visual assessment is effective in delineating benign from the malignant laryngeal lesions, however visual assessment didn't show imaging shape and subject to private variation [15].

In this study diffusion (restricted versus facilitated technique) with sensitivity 91.67% and specificity 75% and accuracy 83.3% with significant P value 0.001 (Table 3) with the area under the curve 0.833 while conventional while conventional MRI show sensitivity of 79.17% and specificity of 50% and accuracy of 64.6% with an insignificant p-value with the area under the curve 0.646. (Table 3).

Our study involves twelve patients with radio necrosis; the lesions show no asymmetric or slight symmetric hyperintesity on DWI (b 1000) and were hyperintense on the ADC map (Fig. 2), the mean ADC was significantly higher than the ADC of the normal laryngeal tissues in the same patients.

Fig. 2.

Male patient, 60yrs old, under follow-up of cancer larynx on radiotherapy. Laryngoscopic examination showed: Thickening of the superior surface of the left vocal cord with no definite mass. Conventional MRI examination (a, b) shows Diffuse thickening of true left vocal cord in T2WI and no focal abnormal signal intensity in STIR sequence. Analysis of ADC value& DWI (c, d) shows: Diffuse faint restricted diffusion of medial aspect of laryngeal air column with ADC value about 1.2 × 10^–3 mm²/s. This is suggesting post-radiotherapy sequalae with no detected residual or recurrent focal massesand this is confirmed by histological examination which revealed no residual malignant cells, post-radiation necrosis

Our study included 24 patients with tumor recurrence. The lesions were hyperintense on DWI (b 1000) and hypointense on the ADC map (Fig. 3), the mean ADC of these lesions was significantly lower than the mean ADC of the normal laryngeal tissues in the same patients, while the mean ADC of the tumor recurrence was significantly lower than the mean ADC of radionecrosis [16, 17].

Fig. 3.

male patient 50yrs old under follow up of on radiotherapy laryngoscopic examination showed large right keratotic necrotic supraglottic soft tissue mass with glottic extension (white arrow), no subglottic extension. It is infiltrating right aryepiglottic fold, right pyriform sinus, right false vocal It is also infiltrating right true vocal cord (A,b)- MRI examination showed hyper intense signal involving right supraglottic region involve areepiglottic fold not separable from the right vocal cord with thyroid and cricoid invasion. Analysis of diffusion & ADC value (c, d) restricted on the previous sites with ADC value 0.7 × 10^–3 mm², histological examination revealed grade II squamous cell carcinoma

Our study consistent with Abdel Razek et al. [18] and Baur A et al. [19], deduce that diffusion-weighted magnetic resonance imaging has the potential to inform apart post-radiation changes from recurrent cancer supported the ADC variations. Abdel Razek et al. [18]. reported mean ADC variations between post-treatment sequels (oedema or radiation necrosis) and residual/recurrent laryngeal cancer; the mean ADC of residual or recurrent lesions (1.17 ± 0.33 × 10^–3 mm²/s) was considerably lower (P < 0.001) than that of post-treatment sequels (oedema or radiation necrosis) (2.07 ± 0.25 × 10^–3 mm²/s.

They explicit that variations in ADC values due to distinct tissue changes, water distribution in tumors, and post-treatment sequels. They added that tissues exhibiting post-treatment sequels, pathological characteristics, such as, tissue with low cellularity associated with variable degrees of edema and rdionecrosis or infection and thus, increase the ADC values [18, 19].

The ADC value of 1.2 × 10^–3 mm²/s was the best cut off value for differentiating tumor recurrence from radiotherapy effect (Fig. 1). This consistent with previous study found that the ADC value of 1.22 × 10^–3 mm²/s or less was the best cut off value for predicting malignancy, with highest accuracy of 86%, with 84% sensitivity, 91% specificity [3], Previous study used an ADC value of 1.30 × 10^–3 mm²/s as the threshold value for differentiating residual or recurrent laryngeal lesions from post-treatment sequel or radionecrosis or infection may resemble malignancy; the best results were obtained with an accuracy of 87%, sensitivity of 84%, and specificity of 90% [18].

Limitations of this study

small sample size due to high cost of imaging; a second is, increase b value (b = 0 and 1000) which may affect measured ADC due to technical problems.; and a third is air artifact in larynx may cause difficult to measures ADC in small lesion. Also ADC maps have poor anatomical delineation and should analyzed with other conventional MRI.

Conclusion

DWI is a non-invasive technique that can be utilised to discriminate tumour recurrence/ residual from postreatment sequels in patients with recurrent /residual lesions after radiotherapy or chemotherapy or concomitant chemo radiotherapy for laryngeal cancer. Because DWI is a quick examination procedure, it can be safely added to conventional MRI to improve diagnostic accuracy.

Abbreviations

MRI

Magnetic Resonance Imaging

ADC

Apparent diffusion coefficient

DWI

Diffusion-weighted imaging

PET

Positron emission tomography

SPIR

Spectral presaturation with inversion recovery

AUC

Area under the curve

ROC

Receiver operating characteristics

STIR

Short-TI Inversion Recovery

SD

Standard deviation

SPSS

Statistical program for social science

NPV

Negative predictive value

PPV

Positive predictive value

Author Contributions

MR,ES,MR: idea of research. MR,ES: referral of patients. AR: collection of data and analysis. MR, SF: imaging, revision and references collection. All authors have read and approved the manuscript.

Funding

None.

Data Availability

The datasets generated during and/or analyzed during the current study are available.

Declarations

Conflict of interest

The authors have no conflict of interest to declare.

Ethical Approval

Ethics approval and consent to participate Ethical committee of Assiut University, Consent was obtained from the study participants before research; informed written consent is available.