Introduction

The lacrimal gland (LG) is known to be a flat orbital structure that produces tears. The acinar portion of the LG is thought to derive from the neuroectoderm [25]. It is superficially located in the lacrimal fossa, a shallow cavity in the superolateral corner of the orbit, and adjacent to the bulb [20]. The LG can be exposed through a lateral orbitotomy [1, 8]. A recent investigation suggested that normal LG—commonly comprising the orbital and palpebral lobes—may exhibit morphological and histological variations [15, 17]. Extensive studies on donated bodies to science, magnetic resonance imaging (MRI), computed tomography (CT), and Doppler sonography have explored normal LGs [2, 4,5,6, 9, 14, 15, 17,18,19, 21, 23,24,25,26,27]. On contrast-enhanced MRI, the LG is delineated as an enhancing area [18, 24]. Lesions of the LG fossa involve diverse pathologies, including inflammation, autoimmune diseases, lymphoproliferative disorders, benign epithelial proliferation, malignant neoplasia, and metastatic disease [12, 22]. Abnormal LG enlargements have been associated with Sjögren’s syndrome, IgG4-related diseases, acromegaly, Graves’ ophthalmopathy, sickle cell disease, and lymphoblastic leukemia [3, 7, 10, 11, 13, 16, 28].

In contrast to investigations of the lacrimal artery (LA), to our knowledge, we miss clear depiction and MRI aspects of the lacrimal vein (LV) [5, 6, 26]. Therefore, the present study aimed to explore the LV and its relationship with the LG using thin-slice contrast-enhanced MRI.

Materials and methods

This retrospective study included 81 outpatients who underwent MRI examinations after presentation to our hospital between July 2010 and May 2015. These patients presented with headaches, dizziness, tinnitus, hearing disturbances, hemisensory disturbances, gait disturbances, scintillation scotoma, and focal seizures. The patient population comprised 39 men and 42 women, with a mean age of 51.1 ± 15.5 years (range, 18–78 years). Patients with a medical history of dry eye, collagen disease, or thyroid orbitopathy were excluded. Initial examinations using axial T1- and T2-weighted sequences, T2-gradient echo, fluid-attenuated inversion recovery, and diffusion-weighted sequences confirmed that the absence of intracranial or intraorbital tumors, cavernous sinus (CS) lesions, symptomatic ventriculomegaly, or traumatic orbital injuries. Subsequently, the patients underwent thin-sliced, volumetric imaging examination with intravenous gadolinium infusion (0.1 mmol/kg) in the axial, coronal, and strictly sagittal planes, involving the entire cranial vault. The following parameters were adopted: repetition time, 4.1 ms; echo time, 1.92 ms; slice thickness, 1 mm; interslice gap, 0 mm; matrix, 320 × 320; field of view, 250 mm; flip angle, 13°; and scan duration, 7 min 25 s. All images were obtained using a 3.0-T MR scanner (Achieva R2.6; Philips Medical Systems, Best, The Netherlands). Imaging data were transferred to a workstation (Virtual Place Lexus 64. 64th edition; AZE, Tokyo, Japan) and independently analyzed by two authors (S.T. and H.I., both of whom had more than 15 years of experience as board-certified neurosurgeons). The LGs and LVs were assessed on the contrast-enhanced axial and sagittal images. On sagittal images, an enhancing, oval or round structure located adjacent to the superolateral surface of the bulb was regarded as the LG. The long (LD) and short diameters (SD) of the LG were measured on a sagittal image, which revealed the largest dimension (Fig. 1). Furthermore, linear intensities connecting the LG to other intraorbital or intracranial venous structures were considered LVs. Because of their low performance in depicting LVs, as confirmed via preliminary observations, coronal images were not used.

Fig. 1
figure 1

Contrast-enhanced sagittal magnetic resonance images showing measurements of the long (b ,d, 1) and short (b, d, 2) diameters of the lacrimal glands (a, c, asterisk) bilaterally

Full size image

The Wilcoxon signed-rank test was used for the statistical analyses. Differences were considered statistically insignificant at p > 0.05.

Furthermore, the LGs and LVs were observed on three dissected donated bodies for science where the bilateral internal jugular veins were manually injected with blue coloring liquid. In two bodies, relationships between the LV and relevant structures were observed in four orbits. In the remaining body, a lateral orbitotomy was performed bilaterally to explore the sagittal dimension of the LG and LV. Dissections were performed by one author (S.T.) at the Department of Neurological Surgery, University of Florida, Gainesville, FL, USA.

This study was conducted in accordance with the guidelines of our institution for Human Research. Written informed consent was obtained from all patients prior to their participation.

Results

Contrast-enhanced MRI examinations

The LGs were consistently delineated bilaterally in all 81 patients, with variable enhancements. On the sagittal images, the mean right LD and SD were measured 17.3 ± 2.4 mm (range: 12.2–24.1) and 13.7 ± 2.1 mm (6.2–18.6), while the left LD and SD were 17.0±2.6 mm (12.0–23.4) and 13.6 ± 2.6 mm (6.7–24.6), respectively (Table 1). Laterality was not observed in the LD or SD groups (p > 0.05). Bilaterally, no sex difference was observed in the LD or SD (Table 2). Furthermore, no specific age range in 10-year increments was associated with a significantly longer LD or SD (p > 0.05). In the sagittal images, two or more vascular structures arose from the superoposterior margin of the LGs. They were identified in 24 of 81 patients (30%), with 9 on the right, 9 on the left, and 6 on both sides (Fig. 2). Due to the low performance in depicting more posterior courses, these vessels were not identified as LVs or LAs. On axial images, the LV was identified in 76 patients (94%). Their course was classified into three types: connecting to the superolateral CS, connecting to the superior ophthalmic vein (SOV), and connecting to the diploic channels of the greater wing of the sphenoid bone (DCGW) (Fig. 3). The CS type was the most frequent and found in 47 of 76 identified patients (62%), with 8 on the right, 5 on the left, and 34 on both sides. The SOV type comprised 22 patients (29%), with 10 on the right, 6 on the left, and 6 on both sides. The DCGW type was the least frequent and observed in 7 (9%) patients, with 0 on the right, 1 on the left, and 6 on both sides (Table 3). The CS, SOV, and DCGW types exhibited highly variable morphologies (Fig. 4).

Table 1 Measured mean diameters of the lacrimal glands
Full size table
Table 2 Laterality and sex difference in measured diameters
Full size table
Fig. 2
figure 2

Contrast-enhanced sagittal magnetic resonance images of three patients showing variable vessels arising from the lacrimal glands on the right (a, dashed arrows), left (b, d ashed arrows), and both sides (c, d, dashed arrows). Asterisk: lacrimal gland. c: right side, d: left side

Full size image
Fig. 3
figure 3

Contrast-enhanced axial magnetic resonance images of different patients showing the lacrimal veins (LVs; (a–c), dashed arrows) with three distinct courses. a LV connecting to the superolateral cavernous sinus, b LV connecting to the superior ophthalmic vein, c LV connecting to the diploic channels of the greater wing of the sphenoid bone Asterisk: lacrimal gland

Full size image
Table 3 Classification of the identified courses of the lacrimal veins
Full size table
Fig. 4
figure 4

Contrast-enhanced axial magnetic resonance images of different patients showing the left lacrimal veins (LVs) with variable courses. (a–c) LV connecting to the superolateral cavernous sinus, (d–f) LV connecting to the superior ophthalmic vein, (g–i) LV connecting to the diploic channels of the greater wing of the sphenoid bone Asterisk: lacrimal gland

Full size image

Donated bodies to science

On dissected specimens, the LVs were consistently coursed between the posterior margin of the LG and the superolateral part of the CS, above the upper margin of the lateral rectus muscle (Fig. 5). Connections between the LV and the SOV were not observed in dissected specimens. When viewed through the lateral orbitotomy, all identified LVs emerged from the posterior margin of the LG and coursed posteriorly along the course of the LAs (Fig. 6).

Fig. 5
figure 5

Posterior and superolateral view of the right orbit on a donated body to science, after the removal of the orbital roof and its underlying periorbita (a) and after the removal of the orbital fat (b), revealing the lacrimal vein (LV) coursing between the posterior margin of the lacrimal gland (asterisk) and superolateral part of the cavernous sinus (CS), above the upper margin of the lateral rectus muscle (LRM). FN frontal nerve, LA lacrimal artery, LM levator muscle, LN lacrimal nerve, SOV superior ophthalmic vein, SRM superior rectus muscle

Full size image
Fig. 6
figure 6

Lateral view of a donated body to science, after a lateral orbitotomy (a) and removal of the periorbita and orbital fat (b), revealing all lacrimal veins (LVs) emerging from the posterior margin of the lacrimal gland (asterisk) and coursing superficially in the posterior direction along the course of the lacrimal artery (LA). LN lacrimal nerve, LRM lateral rectus muscle

Full size image

Discussion

In the present study, the identified LVs on MRIs consistently emerged from the upper posterior margin of the LGs, coursed posteriorly, and drained into the superolateral CS, SOV, and DCGW. Presurgical evaluation of accurate LV courses can contribute to performing function-preserving, safe maneuvers around the LG. Given the intraorbital location and shape of the LGs, sagittal images seemed suitable for delineating the LGs and the LV segments arising from them. However, to our knowledge, no study has explored LGs using sagittal MRI [3, 4, 9, 24, 27]. According to Lee et al. the mean LG width measured on an axial CT was small, being 4.1 mm on the right orbit and 4.3 mm on the left, while that on a coronal CT was 3.6 mm on the right and 3.8 mm on the left [14]. These findings may partly explain why LG data obtained from sagittal MRIs are scant. The present study adopted seamless scanning with a 1-mm slice thickness, which enabled the delineation of the LGs and LVs on sagittal images.

In the present study, laterality or sex differences were not statistically significant between the LD and SD groups. Furthermore, no specific age range was associated with a longer LD or SD. Conversely, a previous investigation using MRI showed that the thickness and area of the LGs decreased with age in women but not in men [27]. The methodology for measuring the LDs and SDs might have influenced the differences because LGs with variable shapes and dimensions are apparently not involved in a sagittal image. Further investigations are required to validate our results.

The present study had several limitations. First, the cohort comprised participants of heterogeneous age and sex. The participants were retrospectively evaluated and not randomly assigned to the MRI examinations. The LVs were assessed only on contrast-enhanced MRI and a small number of donated bodies to science, and the flow patterns of the veins were not assessed. At the present time, we cannot present a rationale for the three different LV types, identified in this study, based on the embryology. Furthermore, on the sagittal images, the LVs were not discriminated from the LAs for low performance in depicting more posterior courses. Despite these limitations, we believe that our results may improve our understanding of the LVs and LGs.

Conclusions

The LV may consistently emerge from the upper posterior margin of the LG. It commonly pours into the SOV or superolateral part of the CS.