Morphological Characteristics of the Lacrimal Apparatus in its Obstruction of Various Genesis

Secondary acquired lacrimal duct obstruction (SALDO) is one of the complications of radioiodine therapy. The material was obtained during endoscopic dacryocystorhinostomy with revision of Hasner’s valve in patients with PANDO (n = 7) in the distal segments of the nasolacrimal duct and in patients with SALDO (n = 7) after radioactive iodine therapy. The material was stained with hemotoxylin and eosin, alcyan blue, and by Masson method. Morphological and morphometric analyses were performed in semi-automatic mode. The results of histochemical staining of sections were translated into points taking into account the area and optical density (chromogenicity). The differences were considered significant at p < 0.05. It was shown that the nasolacrimal duct sclerosis is significantly lower (p = 0.029) in patients with SALDO than in patients with PANDO while fibrosis in the lacrimal sac is the same in patients of the compared groups.


Introduction
The cases of primary acquired nasolacrimal duct obstruction (PANDO) when the reason for its development is unclear are quite common in clinical practice. The cases when the factor that triggered nasolacrimal duct obstruction is known are less common. Such obstructions are called secondary acquired (SALDO). A special case of SALDO is obliteration resulting from radioactive iodine exposure during highdose therapy for differentiated thyroid cancer described for the first time by R.T. Kloos [1].

Clinical Data
The material was obtained during endoscopic dacryocystorhinostomy with revision of Hasner's valve in patients with PANDO (n = 7) in the distal segments of the nasolacrimal duct and in patients with SALDO (n = 7) after radioactive iodine therapy.
In all cases, patients complained of pronounced lacrimation rated 4 points on Munk scale, as well as the periodic mucopurulent lacrimal duct discharge.
All patients underwent a conventional ophthalmological and dacryological examination, as well as CT-dacryocystography, to localize the nasolacrimal duct obstruction.

Obtaining the Material
The samples for histopathological examination were taken from patients with PANDO, who had surgery on the same day as patients with SALDO to minimize the risk of subjectivism when deciding on including the patients in the trial.
After anemization of the nasal mucous membranes and infiltration anesthesia of the inferior nasal concha, it was displaced anteriorly. After eyeball instillation anesthesia, n. infraorbitalis conduction anesthesia and infiltration anesthesia, a probe with visual endoscopic position control was inserted antegradically into the lacrimal ducts of the inferior nasal meatus.
The probe displaced the soft tissues of Hasner's valve anteriorly. In the area where the soft tissue protrusion was the greatest, an incision of Hasner's valve was made using a sickle-shaped scalpel, then a fragment of the nasolacrimal duct wall was removed with Blakesley forceps and sent to the pathohistological laboratory.
Then classical endoscopic dacryocystorhinostomy was performed. Following the opening of the lacrimal sac, protecting the formed flap of the lacrimal sac with the antegrade probe, a fragment of its posterior flap was removed with Blakesley forceps. This fragment was also sent to the pathohistological laboratory.

Morphological Study
Pathological examination was conducted in a specialized laboratory. The samples were previously anonymized and marked with numbers so that a researcher could not identify which group they belonged to. After the examination has been completed and the laboratory findings received, the samples were deanonymized in both groups. The analysis was performed by two skilled histologists. The results, expressed in points, were compared and no differences were found. The results of the qualitative study were the result of consensus.
Fragments of the lacrimal sac and nasolacrimal duct obtained during the surgery were fixed in buffered formalin, processed in automatic mode (Leica Biosystems tissue histological processing apparatus, Germany), the serial Sect. (2 microns thick) were classically stained with hematoxylin and eosin, as well as by Masson method and alcyan blue.
Morphological and morphometric analyses were performed in ten randomly selected fields of the microscope at ×100 and ×400 magnification in four randomly selected sections of each sample, moving the slides at equal intervals along the X and Y axes using a semi-automatic image analyzer. Digital images of histological sections for morphometric studies were obtained using a video microscopy system (Leica DM3000 microscope, Leica DFC450 C camera, Germany), and morphometric data were obtained using Leica Application Suite (LAS) Version 4.9.0 (Germany) image processing and analysis software. To calculate the area of the stromal component, the images were automatically normalized and transferred from the 24-bit color image mode (RGB) to 256 grayscale mode, and binarization of objects with a given sensitivity level was performed. When processing microphotographs by the image analysis system, binary objects with brightness > 150 were cut off and the total area of positively colored objects of the total area of microphotographs was determined. The results of histochemical staining of sections were translated into points taking into account the area and optical density (chromogenicity) in relative units: 1 -weak (0-0.3); 2moderate (0.3-0.6); 3 -significant (> 0.6).

Statistical Analysis
The obtained data were processed using the IBM SPSS Statistics 26 computer program (IBM Analytics, USA). The resulting distribution of values was abnormal according to the Shapiro-Wilk criterion. The nonparametric Mann-Whitney criterion was used for statistical analysis. The differences were considered significant at p < 0.05.

Study Limitations
A limitation of the study was the small size of the samples analyzed. This is due to the objective difficulties in recruiting material. SALDO is a relatively rare pathological condition, and only distal localized obliterations were analyzed in the present study. The quantitative lack of material was compensated for by conducting multiple comparisons as well as by automating the process of analyzing the results of histochemical study. Statistical analysis was performed considering the known limitations.

Results
The average age of patients with PANDO was 57 ± 18 years (40-80 years), and the average age of patients with SALDO was 54 ± 9 years (44-56 years). The gender ratio of women and men was 6:1 in both groups. The average radiological activity administered in patients with SALDO was 114 ± 17 mCi (100-135 mCi), surgical intervention was performed 32 ± 3 months (29-36 months) after radioactive iodine therapy, and 5 ± 6 months (0-13 months) after complaints of lacrimation.
Signs of obliteration and punctate wall ulceration were observed in all samples of the lacrimal sac and nasolacrimal duct obtained from the 1st group of patients (SALDO) after hematoxylin and eosin staining. A significant decrease in the height of the nasolacrimal duct single-layer cylindrical epithelium and autolytic changes in cytoplasm and microvacuolation and partial desquamation were noted ( Fig. 1 A,  Fig. 2 A). Secretory cells of the integumentary epithelium and mucous glands showed different staining on sulfated mucoid substances, hyaluronic acid, and sialomucins which correlated with their morphofunctional state: starting with necrobiotic changes (significant basophilia of cells) and ending with various stages of dystrophic changes, including necrotic leading to lighter staining of cells with a disturbed structure (Figs. 1B and 2B). Most of the acinuses were necrotized, some of them had reduced lumens and abundant microvesicles in the cytoplasm; epithelial cells were hyperchromic. The cell nuclei were protruding, often pyknotic. In some cases, a reversible necrobiotic cells state which was replaced by necrotic changes in the most differentiated and vulnerable cells was noted: pale staining of the cytoplasm due to autolytic processes, as well as fuzzy intercellular boundaries, partial cellular desquamation associated with karyopycnosis and/or karyolysis were observed. A preserved oxyphilic basement membrane was found around the acinar shadows. The stromal component had moderate and severe fibrosis (Fig. 1C), as well as diffuse necrobiosis.
Most of the small-caliber arteries were spasmed, the lumens of the arterioles were narrowed, and significant parts of the hemocapillaries were collapsed and emptied.
Histological examination of biopsies of the lacrimal sac and nasolacrimal duct obtained from the 2nd group of patients (PANDO) revealed keratinization of the integumentary epithelium with squamous metaplasia, as well as epithelial cells with dystrophic changes (Fig. 2 A), revealed after hematoxylin and eosin staining. The cytoplasm of  At the same time, in nasolacrimal duct biopsies of the second group of patients (PANDO), an increase in collagen fibers by 1.5 times (3.0 points) and a decrease in elastic fibers by 3.0 times (1.0 points) compared to the first group (SALDO) in which the collagen-elastic ratio was 2.0:3.0 were observed. (Fig. 2C, Fig. 4C).
The results of quantitative morphometric analysis of the fibrous component of the lacrimal sac and nasolacrimal duct biopsies in patients with SALDO and PANDO are presented in Table 1. When conducting pairwise comparisons, a statistically significant difference (p = 0.029) was found only when comparing the distribution of collagen and elastic fibers in biopsies obtained from Hasner's valve in patients with PANDO and SALDO. Significant differences were also observed when comparing the interaction between elastin and collagen in patients with PANDO and SALDO. glandular cells was stained with mucins and non-sulfated glycosaminoglycans (Fig. 2B). Most acinuses were lined with epithelium with signs of atrophy due to the compression of glandulocytes by surrounding fibrous tissue, denudation, and ischemic changes. Edema, a pronounced fibroplastic reaction with inflammatory infiltration, and areas of fibrosis were observed in the stroma (Fig. 2C). At higher magnification, it was found that the infiltrate consisted of lymphocytes and polymorphonuclear leukocytes which may indicate chronic active inflammation. Vascular congestion was observed.
Histochemical method -Masson's staining -for assessing the fiber component of the intercellular substance revealed a slight increase in collagen fiber bundles (blue staining) and elastic fibers bundles (red-yellow staining) in the lacrimal sacs in the second group of patients (PANDO) compared with the first group (SALDO) representing 3.0 and 2.0 points, respectively (Fig. 1C, Fig. 3C).  The study confirmed the results of a previous study on the relative differences in pathomorphological changes in Hasner's valve in SALDO and PANDO. In addition, statistically significant (p = 0.029) differences in the ratio of collagen and elastin fibers in Hasner's valve in patients with SALDO and PANDO were found which is apparently a consequence of the lower severity of fibrotic processes in patients with SALDO after radioactive iodine exposure than in patients with PANDO. At the same time, similar ratios in the structure of lacrimal sacs, although they were identified, were recognized as statistically unreliable which enables us to characterize the severity of fibrosis in the lacrimal sac as the same in SALDO due to radioactive iodine therapy and in PANDO. The above confirms our hypothesis according to which the primary nasolacrimal duct obstruction after radioactive iodine treatment occurs in the distal parts of the nasolacrimal duct, and the inducing inflammation and fibrotization in the more proximal parts of the lacrimal ducts are secondary. Indirect evidence of these assumptions is the analysis of the anatomical structure of SALDO due to radioactive iodine therapy in the clinical cases described by other authors which, despite the known limitations of such analysis associated with statistically unreliable samples and inaccurate case description, shows the relative predominance of distal forms of such obstruction [5,[7][8][9][10][11][12][13].
It is likely that the qualitative distribution of the stromal component of connective tissue, namely elastin, and collagen, determines some clinical features of the SALDO course. It is fair to assume, for example, that changes in the relationship between these proteins lead to changes in the biomechanical properties of the lacrimal ducts. The change in tissue plasticity, in turn, is connected with its possible atraumatic dilatation, which is relevant during minimally invasive recanalization surgeries, including intubation and balloon dacryoplasty. Our experience shows relatively low clinical efficacy of such surgeries in patients with SALDO, which indirectly confirms this assumption. It is also of particular clinical interest that injury to the nasolacrimal duct due to radiation exposure occurs primarily in its distal parts, while injury of the proximal parts is secondary. Thus, earlier endoscopic plastics of Hasner's valve can be considered as an alternative to dacryocystorhinostomy in more remote terms.

Conclusion
Thus, the comparative morphological study of biopsies of the lacrimal sac and Hasner's valve in patients with SALDO in radioactive iodine treatment and in patients with PANDO not only confirmed the available information regarding the radiation nature of the lesion but also enabled to quantify Discussion According to V. Yartsev et al., SALDO occurs in 8.8% of patients after a single high-dose radioactive iodine therapy, and the probability of the complication is multiplied during repeated treatment [3]. The analysis of the lacrimal sac and nasolacrimal duct biopsies, performed by K. E. Morgenstern et al. revealed the expression of Na + /I − symporter (NIS) in these structures which may be the biochemical basis for the development of SALDO [4]. The modern theory of SALDO pathogenesis due to exposure to radioactive iodine considers this complication a consequence of radioactive iodine in the tear fluid and its active capture by the tear ducts through the opening of the NIS protein channel [5][6][7]. Imaging studies have confirmed that during radioactive iodine therapy fixation of a radioactive drug by the nasal cavity and lacrimal ducts is possible [5,6]. Observations show that SALDO most often develops in the distal parts of the nasolacrimal duct, in particular, at the level of Hasner's valve, although there are cases of more proximal obstruction, including at the level of the lacrimal sac which is accompanied by dacryocystitis.
Previous studies on the comparative pathomorphological analysis of changes in Hasner's valve in PANDO and SALDO due to radioactive iodine therapy have shown that in the latter case specific changes characterized as radiation damage to Hasner's valve develop [2]. It was found that in SALDO desquamation of the nasolacrimal duct epithelium and damage to the mucous glands are observed which is accompanied by moderate fibrosis, while in PANDO fibrosis is relatively more pronounced and is apparently a consequence of a long-running chronic inflammatory process.
We have proposed the hypothesis that the pathological process, primarily induced in the distal segments of the nasolacrimal duct, leads to the obstruction of the lacrimal ducts in situ, and proximal obstruction is a consequence of secondary alteration resulting from excessive reactivity of the tissue. This hypothesis required pathomorphological verification consisting of a comparative study of changes in the lacrimal ducts at various anatomical levels, as well as in the dependence on the cause of the obstruction. Of particular interest is the study of the fibrous component of the lacrimal duct wall, since it was previously shown that it is sclerotic changes that result in SALDO. Such a study has been achieved by Masson's trichrome staining, as well as through the use, in addition to the classical, digital analysis, of the images with a quantitative histochemical evaluation of the results. The alcyan blue staining enabled the characterization of the changes in the mucous glands the role of which in the development of SALDO was explained previously.
of the level of fibrosis of the stromal component and other lacrimal ducts structures. It was shown that the nasolacrimal duct sclerosis is significantly lower (p = 0.029) in patients with SALDO than in patients with PANDO, while fibrosis in the lacrimal sac is the same in patients of the compared groups. This circumstance confirms the hypothesis that the primary lesion when exposed to radioactive iodine affects Hasner's valve, and the alteration of proximal structures is secondary and similar in its pathomorphological characteristics to the process developing with PANDO. The small number of observations in the compared groups can be regarded as a disadvantage of this research. At the same time, the repeatability of quantitative characters was achieved in the groups, which confirms the reproducibility of the study.