The asymmetric facial skin perfusion distribution of Bell’s palsy discovered by laser speckle imaging technology

Abstract

Bell’s palsy is a kind of peripheral neural disease that cause abrupt onset of unilateral facial weakness. In the pathologic study, it was evidenced that ischemia of facial nerve at the affected side of face existed in Bell’s palsy patients. Since the direction of facial nerve blood flow is primarily proximal to distal, facial skin microcirculation would also be affected after the onset of Bell’s palsy. Therefore, monitoring the full area of facial skin microcirculation would help to identify the condition of Bell’s palsy patients. In this study, a non-invasive, real time and full field imaging technology - laser speckle imaging (LSI) technology was applied for measuring facial skin blood perfusion distribution of Bell’s palsy patients. 85 participants with different stage of Bell’s palsy were included. Results showed that Bell’s palsy patients’ facial skin perfusion of affected side was lower than that of the normal side at the region of eyelid, and that the asymmetric distribution of the facial skin perfusion between two sides of eyelid is positively related to the stage of the disease (P < 0.001). During the recovery, the perfusion of affected side of eyelid was increasing to nearly the same with the normal side. This study was a novel application of LSI in evaluating the facial skin perfusion of Bell’s palsy patients, and we discovered that the facial skin blood perfusion could reflect the stage of Bell’s palsy, which suggested that microcirculation should be investigated in patients with this neurological deficit. It was also suggested LSI as potential diagnostic tool for Bell’s palsy.

Keywords

Microcirculation facial skin eyelid laser speckle imaging Bell’s palsy facial paralysis

1 Introduction

Bell’s palsy is one of the most common causes of the abrupt onset of unilateral facial weakness. It is a kind of peripheral facial paralysis due to the problem in the peripheral nervous system [11], and it accounts for 60 to 75 percent of all cases of peripheral facial paralysis [1].

The most probable cause of Bell’s palsy is thought to be virus HSV type 1(HSV-1) [11], which was found in the endoneurial fluid of most Bell’s palsy patients and was absence in patients with the Ramsay Huntsyndrome or other neurologic diseases [18]. It is speculated that HSV-1 reactivation from latency in the geniculate ganglion leads to the outbreak of Bell’s palsy [10]. But how the virus damages the facial nerve is still uncertain. Another longstanding speculation of the cause of Bell’s palsy is ischemia of the facial nerve [5, 11]. It had been reported that occlusion of the artery which is supplying blood to facial nerve could lead to facial paralysis [22]. But most proportion of Bell’s palsy patients’ artery is not occluded. Some investigation had shown that the frequency of diabetes mellitus in Bell’s palsy patients was much higher than that in “healthy” population [17]. No evidence of diabetic neuropathy was existed in these patients, but abnormal blood vessels were frequently showed in their conjunctivae and nail-beds [16]. It had therefore postulated that microangiopathy may be a possible factor for facial nerve lesion. We had also noticed that facial nerve swelling was presence in the reported decompression operations in Bell’s palsy patients [6, 12]. The edema surrounding facial nerve was also evidenced by contrast enhancement magnetic resonance imaging (MRI) after the onset of Bell’s palsy [12, 23]. Edema may be secondary to HSV-1 caused inflammation or ischemia, and will result in elevation of pressure and further vascular damage. Thus, we can speculate that no matter what caused Bell’s palsy, the microcirculation of tissue near facial nerve should most probably be affected after the onset of Bell’s palsy.

The survey of facial nerve blood flow should be helpful to discover whether the microcirculation of tissue near facial nerve was affected by Bell’s palsy. Laser Doppler Flowmetry (LDF) had been used in monitoring rabbit facial nerve blood flow, it was find out acute changes in blood flow with the tympanic segment of the nerve could readily be detected using LDF [20]. But there is not any report of measuring facial nerve blood flow in Bell’s palsy patients till now. In the report of measuring rabbit facial nerve blood flow, the author concluded that the direction of facial nerve blood flow is primarily proximal to distal [20]. Thus, microcirculation of distal tissue, such as skin, could probably be affected if the microcirculation of tissue near facial nerve changes.

In this study, the full area of facial skin blood perfusion of Bell’s palsy patients was measured to discover the influence on facial skin microcirculation by the disease. Participants were divided into different groups based on the House-Brackmann (HB) grade. The HB grading system is contemporarily the standard method for assessing the stage of facial paralysis, it divides facial paralysis into 6 grades from normal to no movement [14]. The relationship between facial skin perfusion and HB grade was analyzed statistically. Long-term measurement was also implemented during recovery.

Laser speckle imaging (LSI) technique was used to measure facial skin perfusion in this study. It is a non-invasive optical based measuring technique for microcirculation perfusion [8], and it has no damage to participants. Compared to traditional optical based measuring techniques such as LDF and Laser Doppler imaging (LDI), LSI is superior in simultaneous high spatial and temporal resolution since it is a full-field imaging technique [3]. When using LSI, the tissue under study is illuminated by a near uniform laser beam, the light scattered by particles in the illuminated tissue forms an interference image (speckle image) and is captured by a camera. These particles fluctuate according to the blood flow, and are integrated into dynamic speckle images of the blood flow when captured by the camera. Higher blood flow will cause more particles movements during integration time of the camera, therefore leads to more blurring of the interference image. By measuring the degree of blurring of interference image, the LSI can give a full-field perfusion distribution image in real-time [3]. LSI devices have not been commercialized until recently, yet had been proved to have excellent intra- and inter- observer reproducibility [15]. However, because of the uncertainties caused by the complexities of light scattering from blood, LSI is only able to give microcirculation blood perfusion values in relative units (e.g. to monitor changes in the perfusion, compared to its previous state or to a calibration blood flow) [3]. For this reason, we used the perfusion measured at the normal side of face for comparison when measuring the change of perfusion at the affected side in this study.

Results showed that Bell’s palsy patients’ facial skin perfusion of affected side was lower than that of the normal side at the region of eyelid, and that the facial skin perfusion difference between two sides of eyelid is positively related to the stage of the disease (P < 0.001). During the recovery, the perfusion of affected side of eyelid increased to nearly the same with the normal side. This study is a novel application of LSI in evaluating the facial skin perfusion of Bell’s palsy, and the results were meaningful to the etiologic and diagnostic study of Bell’s palsy.

2 Materials and methods

2.1 Study participants

The study was approved by the Research Ethics Committee for Shenzhen Institutes of Advanced Technology of Chinese Academy of Science (CAS). Bell’s palsy patients were recruited through Shenzhen Traditional Chinese Medicine hospital. Patients who consented to join the study were sent to the Institute of Biomedical and Health Engineering, Shenzhen Institutes of Advanced Technology, CAS (SIAT-CAS), where facial skin blood perfusion was measured by a LSI device (Normal resolution, PeriCam PSI System, Perimed AB, Sweden). The patients were recruited according to two criteria: 1, age from 18 to 80; 2, unilateral facial-nerve paralysis of no apparent cause, patients with these criteria were excluding: 1, stroke, 2, facial operation, 3, herpes zoster, 4, hemifacial spasm (can’t measured accurately by LSI). Participating patients can quit at any time throughout the study and were treated with ordinary therapy. Healthy participants with no Bell’s palsy was also recruited in the study, they were investigated as Bell’s palsy patients at the stage of HB grade I. The healthy participants were also aging from 18 to 80 and the exclusion criteria of these participants were the same with Bell’s palsy participants.

2.2 Study protocol

The first measurement for participants was assessing with their HB grades, which was conducted by a doctor. All participants’ HB grades were assessed by the same doctor in this study to avoid the subjective bias between doctors.

After the HB grade assessment, the patients were then sent to SIAT-CAS for blinded perfusion measurement. The measurement was conducted under stable ambient temperature (20 ± 1°C). The participants were lying supine quietly for at least 15 min, during which the facial skin blood perfusion was stabilized.

2.3 House-Brackmann grading measurement

In HB grading system, the HB grade score was assessed by observing facial movement in four standard poses: at rest, with a forced smile, with raised eyebrows, and with eyes tightly closed. According to the HB grade score, Bell’s palsy can be divided into 6 grades from normal to no movement [14]. Healthy condition was defined as HB grade I.

2.4 Laser Speckle imaging of blood perfusion

A LSI device (Normal resolution, PeriCam PSI System, Perimed AB, Sweden) was used to carry out the perfusion measurement. The object measured by LSI was illuminated by the laser light wavelength at 785 nm; the reflected light was then collected by a non-contact measurement camera. The degree of blurring was quantified by speckle contrast K, defined as Equation (1), where σ and <I> are the standard deviation and mean intensity of a speckle pattern, respectively. The perfusion was relative to the correlation time τ _c, and τ _c can be calculated through speckle contrast K and integration time T using Equation (2). $K = \frac{σ}{< I >}$ (1) $\frac{σ}{< I >} = {[\frac{τ_{c}}{2 T} {1 - exp (\frac{- 2 T}{τ_{c}})}]}^{\frac{1}{2}}$ (2)τ _c is a relative value which reflects the blood flow [4]. For the commercialized LSI instrument PeriCam PSI used in this study, the perfusion unit was defined as PU. The distribution of PU values generates a color-coded image which reflects the spatial distribution of blood perfusion. In this study, the monitor probe of LSI device was set 25 cm above the people’s face with maximum field of view as 14.8 cm×14.8 cm to ensure the real time imaging of the entire face. The imaging frame rate was 2 fps, the maximum pixel number of each image was 462×344.

2.5 Blood perfusion analysis

Blood perfusion analysis was carried out to firstly find out the region where facial skin perfusion was affected by Bell’s palsy. Since Bell’s palsy is unilateral, and facial nerve blood flow of one side of a face can not be affected by operation in another side [20]. We speculated that microcirculation in the normal side of the face was normal, and microcirculation in the affected side may be different compared to the normal side. On LSI images, we can easily find out the region where the microcirculation perfusion is abnormal at the affected side of face. Secondly, blood perfusion analysis is to reveal the potential correlation between affected blood perfusion value and the stage of Bell’s palsy (the Bell’s palsy patients were categorized by HB grade).

In this study, we define the same area of the two side of the face as the two regions of interest (ROI), mean perfusion value of two ROI in a period of time were calculated and marked as P_n and P_a for the normal side and affected side respectively. A relative perfusion value P_r defined as P_n/P_a was then used in all analysis to eliminated affected factors not related to blood perfusion throughout the study.

2.6 Statistics

At first, we carried out statistical analyses to test whether the difference of P_r value between different HB grade groups is significant. If the distribution of P_r value is normal, F test was used in this study; and if the distribution is not normal, nonparametric test was used [21] (Kruskal Wallis test was used for testing among multiple groups and Mann-Whitney test was used for testing between two groups). The correlation between HB grade and P_r value would also be statistical calculated using partial correlation analysis excluding the influence of age and sex.

The P_r value and HB grade was tracked throughout the patient recovery. Because of dropouts, the latest P_r value and HB grade were regard as the final recovery condition. The difference of P_r value between HB grade groups at the final recovery condition was also analyzed. The difference between first P_r value and final P_r value at same HB grade was analyzed to find out whether the P_r value at a same HB grade varies with time.

3 Results

3.1 Asymmetric blood perfusion distribution exists at the region of eyelid

The distribution of dermal perfusion on Bell’s palsy patient’s face was observed asymmetrically. Apparent difference was observed between affected side and normal side of eyelid, which were showed on the color-coded PSI images (Fig. 1a), on almost every Bell’s palsy patients. The facial skin perfusion was different significantly between two sides of eyelid (Fig. 1b). The perfusion was not different significantly on the other part of symmetrical regions. There were also no such apparent difference on HB grade I group (healthy people). There were also no significant differences between the two sides of face in regions except eyelid. So the region of eyelid was an ideal area to monitor the perfusion differences between two sides of face. Statistical analysis was based on the P_r value of the regions of eyelid, as shown on Fig. 1a.

3.2 Significant difference of P_r value is relative to HB grades at initial assessment

Of 85 people who participated in the study, 10 were healthy (HB grade I), 75 were Bell’s palsy patients (HB grade from II to IV). Of these patients, 32 were assessed as HB grade II, 31 were HB grade III, and 12 were HB grade IV. The mean value and standard deviation of P_r of different HB grade groups was depicted on Fig. 2 and Table 1. P_r value of Bell’s palsy was larger than 1(1.25 ± 0.18) significantly (P < 0.001) in contrast to healthy people’s value near 1(1.00 ± 0.06). With higher grades indicating worse facial paralysis from II to IV, the mean P_r values increased significantly (II, 1.13 ± 0.07; III, 1.24 ± 0.08; IV, 1.57 ± 0.19; P < 0.001). The partial correlation analysis was used to test the correlation between HB grades and P_r value, excluding the influence of age and sex. The partial correlation coefficient of HB grade and P_r value was 0.827 (P < 0.001). Take P_n as the value of normal tissue, the relative value P_r would near 1 if P_a was nearly the same with P_n. In Bell’s palsy patients, P_r value was larger than 1, which indicated that the perfusion of affected side was lower than that of the normal side. The results were corresponding to our speculation that facial skin perfusion of affected side would be influenced by Bell’s palsy. And it seems that higher grade of Bell’s palsy has larger impact on facial skin perfusion of affected side.

3.3 Significant difference of P_r value still existed among different HB grades at final measurement

We had further studied the influence of Bell’s palsy on facial skin perfusion during recovery. Since the time course of continuous measurement during recovery was not the same among patients, we only analyzed the final recovery state of Bell’s palsy patients statistically. Patients with more than one measurement were included in the recovery research, so 25 patients were dropped out because they only take measurement once. There were 50 patients left in the research; and at the final recovery state, 11 of them were assessed as recovering to grade I, 19 of them to grade II, 9 of them to grade III, 2 of them to grade IV. The distribution of P_r values among different HB grade groups (excluding group IV) was still significant (P < 0.001) at the final recovery state, which was depicted on Table 2. The group IV was excluded because only 2 examples were included, and it has no statistical meaning. Similar to the trend of the initial measurement, P_r value was higher at worse Bell’s palsy stage. It should be noticed that 9 patients were found especially different from others and could not be categorized to the HB grade. These patients had P_r value less than 1 significantly (P < 0.001), which indicated the perfusion of affected side of eyelid was higher than the normal side. Doctor’s diagnosis of these patients was nearly fully recovery, but between HB grade II and HB grade I. Therefore, we classified these patients as Reverse group, which means the facial skin perfusion of the affected side over recovered to be higher than the normal side.

To test the consistency between initial measurement and final measurement, P_r values of same HB grade but different measurement time were analyzed statistically. As shown on Table 2, at the confidence level of 99% , the difference between initial and final measurement at HB grade I (P = 0.181), HB grade II (P = 0.039) and HB grade III (P = 0.197) were not significant.

3.4 The time course of P_r value in the process of recovery

In Bell’s palsy patients with continuous measurements in the process of recovery, we had found that the P_r value decreased to nearly 1 gradually. An example of this time course of P_r value is shown on Fig. 3. It should also be noticed in Fig. 3 that the P_r value increased if the condition deteriorated, which means the P_r value was incidental to the condition of Bell’s palsy patient. As the condition of Bell’s palsy varied, the P_r value varied accordingly. Higher P_r value was related to worse condition, whereas lower P_r value was related to better condition.

4 Discussion and conclusion

In this study, the asymmetric facial skin perfusion distribution of Bell’s palsy patients was discovered by LSI. The region of eyelid was found the most obvious area to show the asymmetry of facial skin perfusion distribution. The perfusion of affected side of eyelid was lower than that of the normal side in Bell’s palsy patients whereas the perfusion of the two sides was nearly the same in healthy participants. The results evidenced that facial skin perfusion was affected after the onset of Bell’s palsy. As the condition deteriorated from HB grade I to IV, the difference between normal side and affected side of eyelid became larger (patients of higher HB grade had higher Pr values). To find out the influence of Bell’s palsy on facial skin perfusion during recovery, continuous measurements were carried out and final recovery state was analyzed statistically. Both the time course of continuous measurements in the process of recovery and the statistical analysis of final measurement had shown that the perfusion condition changed as the stage of Bell’s palsy varied. It is noticeable according to our study that superficial facial skin perfusion could effectively reflect the stage of Bell’s palsy.

A possible reason for the connection between facial skin microcirculation and Bell’s palsy may because of the edema surrounding facial nerve, which was well evidenced by decompression operation and MRI in Bell’s palsy patients [12]. The edema will result in elevation of pressure and further vascular damage surrounding facial nerve [17]. As the direction of facial nerve blood flow is primarily proximal to distal [20], microcirculation of distal tissue, such as skin, could probably be affected if the microcirculation of tissue near facial nerve changes. As the further damage caused by edema surrounding facial nerve developed, the lesion of facial nerve could be more serious; and the influenced microcirculation of facial skin could be lower accordingly. In the process of recovery, the edema is eliminated gradually and the microcirculation surrounding facial nerve is recovery, and the influenced facial skin perfusion could accordingly increase to normal condition with time. The over recovery of facial skin perfusion may occur before the fully recovery of the lesion, which was observed in this study (P_r value decreased to less than 1 before fully recovery).

But there are still some uncertainties of this study. The most curious one is why the asymmetry only existed in the region of eyelid. A possible explanation of this phenomenon may lie on the thickness of facial epidermis and the penetration depth of LSI. The main nerve trunk was 20.1 ± 3.1 mm [19], and supraorbital nerve superficial branches ranges from 2 to 6 mm [7], but the penetration depth of LSI is hardly to reach depth more than 1 mm [9]. Actually, in most part of facial skin, epidermis was more than 1 mm except eyelid [2, 13]. Since dermal microcirculation of human being is mainly under the layer of epidermis, LSI instrument used in this study can not even sample the dermal microcirculation at the region except eyelid. If the penetration depth of LSI could reach to dermal layer in most part of facial skin, we speculated that asymmetric perfusion distribution between normal and affected side could be found on almost the entire face.

In conclusion, although the reason why facial skin perfusion is affected by Bell’s palsy needs further investigation, the facial skin perfusion changes were discovered firstly after the onset of Bell’s palsy in this study. To the best of our knowledge, it is the first work on monitoring facial skin perfusion of Bell’s palsy patients. The results suggest a novel approach to study this neural disease in the light of microcirculation. Besides, the perfusion measured by Laser speckle imaging (LSI) instrument could reflect the stage of Bell’s palsy. It was a novel and successful demonstration of LSI technique in grading the Bell’s palsy, which suggests the potential of LSI on diagnosis of Bell’s palsy.

Footnotes

Acknowledgments

This work was supported by the Shenzhen Science and Technology Innovation Committee grants: CXB201104220030A. The authors would like to acknowledge Jun Liu and Bin Yang for amending the English expression of the manuscript.

References

Adour

K.K.

, Byl

F.M.

, Hilsinger

R.L.

, Kahnand

Z.M.

and Sheldon

M.I.

, The true nature of Bell’s palsy: Analysis of 1,000 consecutive patients, Laryngoscope 88 (1978), 787–801.

Barker

D.E.

, Skin thickness in the human, Plast Reconstr Surg 7 (1951), 115–116.

Briers

J.D.

, Laser Doppler, speckle and related techniques for blood perfusion mapping and imaging, Physiol Meas 22 (2001), R35–R66.

Briers

J.D.

and Webster

, Laser speckle contrast analysis (LASCA): A nonscanning, full-field technique for monitoring capillary blood flow, J Biomed Opt 1 (1996), 174–179.

Calcaterra

T.C.

, Rand

R.W.

and Bentson

J.R.

, Ischemic paralysis of the facial nerve: A possible etiologic factor in Bell’s palsy, Laryngoscope 86 (1976), 92–97.

Cawthorne

, The pathology and surgical treatment of Bell’s palsy, Proc R Soc Med 44 (1951), 565–572.

Christensen

, Lachman

, W

W.P.

and CL

C.B.

, Cutaneous depth of the supraorbital nerve: A cadaveric anatomic study with clinical applications to dermatology, Dermatol Surg 40 (2014), 1342–1348.

Forrester

K.R.

, Tulip

, Leonard

, Stewart

and Bray

R.C.

, A laser speckle imaging technique for measuring tissue perfusion, IEEE Trans Biomed Eng 51 (2004), 2074–2084.

Fredriksson

, Larsson

and Stromberg

, Measurement depth and volume in laser Doppler flowmetry, Microvascular Research 78 (2009), 4–13.

10.

Furuta

, Takasu

, Sato

K.C.

, Fukuda

, Inuyama

and Nagashima

, Latent herpes simplex virus type 1 in human geniculate ganglia, Acta Neuropathol (Berl) 84 (1992), 39–44.

11.

Gilden

D.H.

, Bell’s Palsy, N Engl J Med 351 (2004), 1323–1331.

12.

Gilden

D.H.

and Tyler

K.L.

, Bell’s palsy is glucocorticoid treatment enough, N Engl J Med 357 (2007), 1653–1655.

13.

R.Y.

, Nojima

, Adams

W.P.

and Brown

S.A.

, Analysis of facial skinthickness: Defining the relative thickness index, Plast Reconstr Surg 115 (2005), 1769–1773.

14.

House

J.W.

and Brackmann

D.E.

, Facial nerve grading system, Otolaryngol Head Neck Surg 93 (1985).

15.

Humeau-Heurtier

, Abraham

, Durand

and Mahe

, Excellent inter- and intra-observer reproducibility of microvascular tests using laser speckle contrast imaging, Clin Hemorheol Microcirc 58 (2014), 439–446.

16.

Korczyn

A.D.

, Aetiology of Bell’s palsy, BRIT MED J 4 (1971), 491–492.

17.

Korczyn

A.D.

, Bell’s palsy and diabetes mellitus, Lancet 297 (1971), 108–110.

18.

Murakami

, Mizobuchi

, Nakashiro

, Doi

, Hato

and Yanagihara

, Bell palsy and herpes simplex virus: Identification of viral DNA in endoneurial fluid and muscle, Ann Intern Med 124 (1996), 27–30.

19.

Rudolph

, Depth of the facial nerve in face lift dissections, Plast Reconstr Surg 85 (1990), 537–544.

20.

Sillman

J.S.

, Levine

R.A.

and Kobler

J.B.

, Laser doppler measurements of intratemporal facial-nerve blood flow, AM J OTOL 15 (1994), 327–334.

21.

Sun

, Medical Statistics(Third Edition), People’s Medical Publishing House, (2002).

22.

Vries

N.D.

, Versluis

R.J.

, Valkand

and Snow

G.B.

, Facial nerve paralysis following embolization for severe epistaxis, J Laryngol Otol 100 (1986), 207–210.

23.

Yetiser

, Kazkayas

, Altinok

and Karadeniz

, Magnetic resonance imaging of the intratemporal facial nerve in idiopathic peripheral facial palsy, Clin Imaging 27 (2003), 77–81.