Are textbook lungs really normal? A cadaveric study on the anatomical and clinical importance of variations in the major lung fissures, and the incomplete right horizontal fissure

The lungs have three main fissures: the right oblique fissure (ROF), right horizontal fissure (RHF), and left oblique fissure (LOF). These can be complete, incomplete or absent; quantifying the degree of completeness of these fissures is novel. Standard textbooks often refer to the fissures as complete, but awareness of variation is essential in thoracic surgery.


| INTRODUCTION
In anatomy textbooks, the lungs are typically described as being divided into lobes by three main fissures. The right lung is divided into upper, middle and lower lobes by the oblique (ROF) and horizontal (RHF) fissures, while the upper and lower lobes of the left lung are separated by the oblique fissure (LOF). The lobes are formed during embryonic development through obliteration of the spaces between bronchopulmonary segments, except at the sites of the three main fissures. Fissures comprise a double-layer of invagination of the visceral pleura, which separates the lobes of the lung. These areas may become partially or totally obliterated, resulting in an incomplete or absent fissure, respectively.
In addition to the main fissures, accessory fissures may also be present. In the right lung, the apicomedial part of the upper lobe is separated from the rest of the right upper lobe by a visible azygos fissure, which contains the azygos vein. This accessory fissure is thought to form due to anomalous embryological development of the azygos vein (Akhtar, Lal, Martin, & Popkin, 2018). Rarer variants on the right include, an inferior accessory fissure separating the medial basal segment from the rest of the lower lobe, and a superior accessory fissure which divides the apical segment of the lower lobe from the basal segments. In the left lung, a horizontal fissure may occasionally be found (Gatzoulis, 2008).
These anatomical variations have implications in thoracic surgery.
The presence of incomplete fissures in patients with early stage adenocarcinoma of the bronchus has been shown to correlate closely with poorer long-term outcomes after attempted curative resections (Lee, Lee, Lee, Kim, & Chung, 2016;Okamoto, Kubokura, & Usuda, 2018). In video-assisted thoracoscopic lobectomy for early-stage nonsmall-cell lung cancer, the degree of pulmonary fissure completeness is likely to result in more severe postoperative cardiopulmonary complications ; in addition, the degree of completeness is an excellent categorical predictor for both major and minor morbidity in video-assisted thoracoscopic lung cancer lobectomy . Following any lobectomy, the presence of an incomplete fissure is a discrete risk factor for a prolonged air leak, leading to extended hospitalization for chest drainage (Stamenovic, Bostanci, Messerschmidt, Jahn, & Schneider, 2016). Incomplete fissures can also result in the failure of bronchoscopic lung volume reduction surgery, permitting collateral ventilation between lobes (Koster & Slebos, 2016). In addition, an absent or incomplete main fissure can cause spread of infective disease such as bronchopneumonia between adjacent lobes (Tarver, 1995). In summary, the less complete the major fissures are, the worse the patient outcomes.
In diagnostic radiology, lung fissures constitute important landmarks in chest radiographs and computed tomography (CT) scans.
Alterations in the appearance of the main fissures can aid in the diagnosis and location of a number of pathologies, including atelectasis, pleural effusions and mass lesions; a well-known example is a pleural effusion tracking through the RHF giving an appearance of a horizontal white band on chest radiographs (Hayashi et al., 2001). Anatomical variants within these fissures can therefore lead to diagnostic difficulties. Finally, it is also important for respiratory clinicians to recognize the existence of accessory fissures in segmental localization of pulmonary lesions or in the assessment of the extent of pulmonary disease, as they may be misinterpreted as pathological structures (Yildiz et al., 2004).
Variations in incomplete or absent lung fissures are well reported in cadaveric and radiological papers, however the extent of this incompleteness has not previously been described. This observational study aims to determine the frequency, as well as the extent of completeness of the main fissures in cadaveric lungs to highlight their importance to thoracic surgeons to aid in operative planning. The number and location of accessory fissures found will also be described.

| MATERIALS AND METHODS
The sample consisted of 82 formalin embalmed cadavers, which were donated over two academic years. These comprised of 42 females and 40 males, with an average age at death of 84.8 (range 63-100).
All subjects were white Caucasians from the United Kingdom, and none had undergone pulmonary surgery. One individual was excluded as they had advanced metastatic pulmonary disease, which precluded detailed assessment of the lung fissures. A total of 81 pairs of lungs were examined as part of the study. Patients with benign respiratory pathology were included as it was clear where the fissures lay in these cases.
For each pair of lungs the main fissures were noted as being complete, incomplete or absent (see Figures 1-3, respectively). When present, each ROF and LOF was measured from the lung hilum anteriorly to the hilum posteriorly. Measurements were taken along the ROF and LOF, as per the respective red and blue lines demonstrated in For an incomplete fissure, both its actual length and its projected complete length were recorded. The projected complete length was an estimate of where an incomplete fissure would continue, by assuming a continuation along a similar contour to their end point were they to be complete. The degree of completeness was then expressed as a percentage of the total projected length (see Figure 2), that is, a complete fissure would have a percentage of 100%. The mean percentage completeness for each main fissure was calculated. The SD on these values was then performed to project how far a main fissure would extend before its likely interruption by lung parenchyma. By expressing the fissures as these percentages rather than absolute measurements variations due to body size, sex or differing states of lung inflation upon embalming would be reduced. The absolute values for mean lengths were also reported for to allow for comparison with the literature. The number of accessory fissures present in each lung was recorded and in one academic year (41 cadavers) the location of the accessory fissure was recorded. Data analysis was performed using Microsoft Excel to calculate SD and percentages. All donors had consented for anatomical research and non-identifiable use of anatomical images prior to death in accordance with the Human Tissue Act (2004).

| RESULTS
In 81 left lungs, 13 incomplete (16.0%) and two (2.47%) absent LOF were found. In 81 right lungs, 29 (35.8%) incomplete and zero absent ROF were found; 54 (66.7%) incomplete and nine (11.1%) absent RHF were observed. The mean length of the LOF was 20.8 cm (SD 4.85) and mean percentage of completeness was 97.1% (SD 8.03). The mean ROF length was 19.3 cm (SD 4.97) and mean percentage of F I G U R E 2 The right horizontal fissure extends 65 mm from the right oblique fissure before it is interrupted. The dashed line denotes the continuation the right horizontal fissure were it to be complete. In this case, the right horizontal fissure is stated to be 54.5% complete [Color figure can be viewed at wileyonlinelibrary.com] F I G U R E 3 The right horizontal fissure is completely absent [Color figure can be viewed at wileyonlinelibrary.com] completeness was 91.6% (SD 15.3). The mean RHF was 9.25 cm (SD 3.88) and mean percentage of completeness was 69.4% (SD 23.8). Table 1 summarizes these findings.
All incomplete ROF and LOF were deficient posteriorly at the lung hila. When present the RHF always had its origin from the ROF and ran towards the anteromedial lung edge, if incomplete the deficiency would be between where the fissure finished and the anteromedial lung edge (see Figure 4). Accessory fissures were present in 10 (12.3%) left lungs and 19 (23.4%) right lungs. In the 41 cadavers where the nature of accessory fissures was recorded there were five left upper lobe horizontal fissures, two azygos fissures, four RHF that continued onto the right lower lobe and one right lower lobe inferior fissure.
Overall, eight (9.88%) right lungs and 58 (71.6%) left lungs from the total 162 lungs studied had a textbook appearance, with complete main fissures and no accessory fissures. The frequency of such a "typical" lung in this study was therefore 40.7%.
This study is the single largest using formalin fixed cadavers and the largest on a population from both the United Kingdom and Europe, it is also the only paper that assesses the extent of completeness of incomplete fissures. The anatomy of the lung fissures varies considerably between these studies. These differences may be attributable to heterogeneity between methodologies and, possibly, ethnic variation.
Tables 2-4 summarize the findings of the literature for the LOF, ROF and RHF respectively. The current study and previous studies demonstrate there is an overall preponderance for the LOF to be complete, with at least 50% of LOF studied being complete. Dutta et al. (2013) report and Kc et al. (2018) Lukose et al., 1999;Magadum et al., 2015;Murlimanju et al., 2012;Nene et al., 2011;Quadros et al., 2014;Unver Dogan et al., 2015;Varalakshmi et al., 2014;Wahane & Satpute, 2015). The RHF is also the most likely major fissure to be absent and in this study 11.1% of RHF were not present. Ghosh et al. (2013) and Kaul et al. (2014) report the highest rates demonstrating absence of RHF in 47.8% and 40% of right lungs respectively. Only Mamatha et al. (2016) found no absence of RHF in their study. The RHF is therefore the most variable of the major fissures, and due to this will exert a large clinical influence due to these higher levels of unpredictability.
Variable findings have also been reported in studies of radiological images. It is difficult to assess the completeness of lung fissures on chest radiographs, as only the RHF can be reliably observed (Suwatanapongched et al., 2015).  Tables 1-4, other studies, however, have demonstrated that pre-operative CT scan reports are not an accurate predictor of the main fissure anatomy found during thoracic surgery (Kent, Ridge, & O'Dell, 2015;Schieman et al., 2011).
There is uncertainty as to whether cadaveric studies overestimate incomplete lung fissures, as the populations studied are more likely to be elderly individuals where fusion may have occurred during life. One study has been carried out looking at the lung fissures in fetal specimens, where incomplete fissures were reported in the ROF and RHF (Zareena, 2014). Where specimens are fixed in formalin again there is a risk this process can result in obliteration of tissue planes, causing an increased reporting of incomplete fissures.
This was not the experience of the study conducted here, and in autopsy studies, where no embalming had occurred, incomplete and absent fissures are demonstrated (Dhanalakshmi et al., 2016;Medlar, 1947;Unver Dogan et al., 2015). The anatomy of the fissures found in cadaveric studies should therefore be more representative of what thoracic clinicians can expect to find at the time of surgery, in particular in comparison to radiological studies investigating lung fissure anatomy.
T A B L E 4 Comparison of incomplete or absent right horizontal fissures (RHF) within the literature. Population of origin and types of specimen used are listed-formalin fixed cadavers (FFC), unspecified cadavers (UC) and autopsy study (AS). Where referenced lengths in centimeters (cm) and standard deviations (SD) are given Absolute measurements of lengths of fissures were reported in other studies, but the usefulness of this data is debatable in cadaveric lungs, the most important reason being that the level of inflation at the time of study varies from complete collapse to full inflation. In life, at the time of thoracic surgery, the lung fissures will again vary in length with inhalation and exhalation. This is the first study where the percentage and degree of completeness of fissures has been calculated. This is important as its value should remain relatively constant despite the changes pointed out above.
A good understanding of the anatomy of pulmonary lobes and fissures allows the surgeon to correlate imaging with the pathological process, and ensures the best possible operation for each patient (Ugalde, Camargo, & Deslauriers, 2007). This study demonstrates that in an elderly UK population the LOF and ROF tend to be near complete (97.1% and 91.6%, respectively) and are only deficient to a small degree posteriorly. Of note Dutta et al. (2013) also demonstrated LOF and ROF that were incomplete anteriorly. Thus, in cases where lung resection is required, the surgeon should carefully examine the specific areas where the oblique fissures approach the lung hila to ensure that an incomplete fissure is stapled off adequately in order to avoid a persistent air leak. During bronchoscopic lung volume reduction surgery again these areas should be checked to ensure that there is no collateral ventilation between lobes. If feasible incomplete lung fissures can be stapled off, or if the fissure is not large enough the procedure may have to be aborted or revised. When oncological resection is being performed and an operable tumor involves a segment of an incomplete LOF or ROF close to where the lobes are fused, an additional lobectomy or total pneumonectomy may be required for optimal tumor clearance.
Where the fissures are largely incomplete, patients should be told perioperatively that they may require an extended hospital stay for chest drainage due to the potential for persistent air leaks.
The anatomy of the RHF is much more variable, with only a minority being complete. Surgeons should be aware that it will likely be incomplete or absent, and have a lower threshold for careful examination and taking the steps described above to ensure good outcomes in interventional procedures. They can expect the RHF to pass approximately 69.4% (SD 23.8) of the distance from the ROF to the anteromedial lung edge, on the basis of this study in an elderly UK population. Craig and Walker (1997)

| CONCLUSIONS
In conclusion, this study not only confirms what was known from studies in different populations, but in addition demonstrates this variability in the largest cadaveric European cohort in the literature. It is also novel in its approach to characterizing the percentage completeness of the major fissures of the lungs. The variability in completeness of the RHF, in particular, was considerable and in the population studied this fissure was 69.4% (SD 23.8) complete.
For thoracic surgeons this has an important impact on oncological and benign procedures, as well as post-operative complications.
This study provides an estimate of the extent of incompleteness that can be expected at operation, to ensure patients receive the most appropriate operation performed optimally. Radiologists and respiratory physicians reviewing radiological studies should also be aware of these gross anatomical variations, and that they have an important influence on diagnostic imaging and respiratory pathology.
A typical "textbook" lung has three complete major fissures and no accessory fissures. In this study only 40.7% of had this classical appearance. Clinicians and anatomists should be aware that lung fissure anatomy is more variable than traditionally presented.

ACKNOWLEDGMENTS
We acknowledge the generosity of the donors who made this study possible, and the invaluable assistance provided by Maria Wright, Senior Teaching Technician, and Darren Broadhurst, assistant teaching technician, in the Human Dissection Room.