Endophytic Tumour Area: A New Morphometric Marker That Predicts the Outcome of Laparoscopic Partial Nephrectomy


 Aims:To determine morphometric and clinical variables associated with ischaemia time (IT) and complications in laparoscopic partial nephrectomy (LPN).Material and methods:Retrospective observational cohort study of 210 patients treated with LPN between 2006-2019.The new morphometric variables that we proposed, were: total tumour area (TTA), line of contact (LC), endophytic tumour area (ETA), exophytic tumour area (XTA). Central tendency and dispersion measures were calculated.Patients were divided into two groups according to IT (IT<15min(n=99) and IT≥15min(n=111)), for the purpose of statistical analysis. Univariate analysis was performed using the Mann-Whitney U test, and multivariate logistic regression analysis to identify variables associated with IT and complications, and COR curves and their area under curve (AUC) of ETA to predict IT and complications.Results:Mean age was 66.62±13.33 years (28-90).There was correlation between IT and ETA (B: 0.09; p<0.001), and correlation between IT and experience (B: -0.07; p<0.001).Variables that significantly modified IT, were between others: ETA (p<0.001) or LC (p=0.001). ETA (p=0.037) and LC (p=0.044), were associated with the occurrence of complications.The multivariate analysis to predict IT, showed that experience (p<0.01) and ETA (p=0.001) independently predicted IT; a second multivariate analysis to predict global complications, showed that AC (p=0.005), and TTA (p=0.024) independently predicted global complications. A third multivariate analysis to predict major complications (Clavien >=III), showed that experience (p=0.021), TTA (p=0.037), and Charlson (p=0.039), predicted major complications.The COR AUC of ETA was 0.730 to predict IT; 0.616 to predict overall complications, and 0.678 to predict major complications.ConclusionsETA and surgical experience are independent variables associated with IT in LPN.CL and TTA are independently associated with overall complications. TTA, Charlson index, and experience independently predict major complications.


Introduction
Kidney cancer is the third most frequent urological tumour, after prostate and bladder cancer.
According to data from the SEOM (1) (Spanish Medical Oncology Association), in 2018 and in terms of prevalence, kidney cancer ranked fourteenth among the most frequent neoplasias at worldwide level, with a total of 403,232 cases diagnosed (2.2%). In this same year (2018), and according to the same source, in Spain renal neoplasia was in seventh place in the national tumour ranking (with 21,584 cases (2.8%)).
As regards the death rate attributable to kidney cancer, and according to the Spanish National Institute of Statistics (INE) (3), the number of deaths in 2018 was 2,131, of which 1,414 were men.
In epidemiological terms, as established in the international consensus guide of the EUA (European Urology Association) (4), it should be noted that in terms of gender, kidney cancer is more frequent in men; it is more frequent in African-Americans, and the majority of cases are diagnosed between the fourth and sixth decade of life. The most signi cant risk factors are smoking, obesity, and HTA. Clinically, patients are usually asymptomatic, and incidental diagnostic involve 50% of cases. (4) Regarding its therapeutic management, this depends on several factors, including the size of the tumour, presence of metastasis, or clinical status of the patient, etc. In this case, renal surgery with curative intention, partial nephrectomy (PN), is the treatment of choice for tumours in stage cT1 (2017 TNM classi cation), which strongly recommend PN (7) (8) (9) Given the complexity of partial nephrectomy, which in most cases requires prior planning, we show through this study how certain morphometric characteristics of the renal tumour not only help to plan the intervention, but can even predict the surgical results, expressed in terms of ischaemia time and complications, two of the major concerns faced by the urologist during and after surgery.

Materials And Methods
We carried out a retrospective observational cohort study of 210 patients with renal tumours, who were treated by means of laparoscopic partial nephrectomy (LPN), performed by a single surgeon (RAA), between 2006 and 2019 in the Ferrol Healthcare Area.
None of the patients presented any formal contraindication, neither clinical nor anaesthetic, that would impede the surgery.
All patients, sign an informed consent.
Using the application of the image viewer of our centre, we describe the following variables related to the morphology of the tumour (proposed morphometric variables, all measured at the point of maximum diameter of the tumour): Endophytic Tumour Area (ETA), de ned as the tumor area/surface within renal parenchima, and expressed in cm²; Exophytic Tumour Area (XTA), de ned as the area/surface of the tumour that protrudes from the renal surface/outer layer, expressed in cm²; Total Tumour Area (TTA), as the total area occupied by the tumour, resulting from the sum of ETA and XTA, and expressed in cm²; and Line of Contact (LC), only applicable to tumours with an exophytic component, and de ned as the diameter of the tumour at the time it passes through the surface of renal parenchyma, and expressed in cm.
For the quanti cation of the morphometric variables that we propose, we used the application of the image viewer of our centre, which makes it possible, by tracing points and lines, to calculate areas (in mm²), as well as the distance between two points (in mm). Figures 1 and 2 We also analysed demographic variables (sex and age, among others), classic morphometric variables (PADUA and RENAL), clinical variables (Charlson comorbidity index and perioperative complications), and experience (as a discrete quantitative variable), referred to a single surgeon and graduated in ascending order (so that 1 ( rst LPN), would correspond to minimum experience, and 210 (last LPN in our series), to maximum experience).
Regarding complications, the Clavien-Dindo rating scale was used, de ning as major complications those of grade ≥ III.

Statistical analysis:
Central tendency and dispersion measures were calculated. The Mann-Whitney U-test was used for the analysis of differences between groups. Multivariate logistic regression analysis was performed to identify variables associated with ischaemia time (IT), and complications; the ischaemia time used as a cut-off point was 15 minutes, which corresponds to the median of the IT of the sample, thereby categorising, by statistical interest, the patients into two groups: >15' and ≤15' COR curves and their area under the curve (AUC) were also performed to determine the ability of the endophytic tumour area (ETA) to predict IT and complications (global and major).
A linear regression analysis was also performed to determine the association of proposed morphometric variables with IT.
A value of p <0.05 was considered statistically signi cant.
The Statistical Package for the Social Sciences (SPSS®), version 22 (SPSS Inc., IBM Corp., Armonk, NY, USA) was used for the analytical calculations.
Measurement of proposed morphometric variables: Using a CT image (axial view), and by using the applications of the image viewer, a dotted, curved, and/or straight line is drawn, which makes it possible to calculate areas (expressed in mm²) and/or distances between two points (expressed in mm).

Results
The data, in terms of demographic, clinical and morphometric variables, is shown in Tables 1, 2 and 3.
No reconversion to open surgery was required in any case; 4 patients (1.90%) required re-intervention after surgery, in all cases due to hemorrhagic complications.
The most frequent complications were the need for transfusion due to anaemia (14 patients, 6.66%) and perirenal haematoma (4 patients, 1.9%), and all of them were resolved with the use of conservative measures.
According to histological type tumour, clear cell carcinoma was the most frequent, 112 (53.33%).
Five patients (2.38%) presented tumour recurrence; metastases developed during follow-up, and as a result of renal neoplasia, four patients (1.90%).
Using linear regression analysis, a signi cant and positive correlation was found between IT and ETA (B: 0.444; p<0.001), and a signi cant and negative correlation between IT and experience (B: -0-542; p<0.001).
Multivariate logistic regression analysis was performed to identify variables associated with ischaemia time (IT), and surgical complications (overall and major). Table 4 & 5.
Lastly, the area under curve (AUC) COR of the endophytic tumour area was 0.730 to predict IT, 0.616 to predict overall complications, and 0.678 to predict major complications.

Discussion
Based on the different types of surgical approaches used to carry out a PN (laparoscopic, robotic, or open), numerous studies have focused on ascertaining which of them is the safest in terms of the oncological, clinical, or surgical results, without any statistically signi cant differences having been found between them (5) (6).
With regard to experience in laparoscopy, Porpiglia et al (10), show in their study that for the same surgeon, as the number of LPN carried out increases, and therefore their experience, there is a decrease in the number of complications, IT, and the positivity of the surgical margins. These results have been con rmed on several studies, not only in the eld of urology, but also in many other elds. Once a certain minimum experience has been achieved, a laparoscopic surgery is not inferior to other types of surgical The purpose of Porpiglia`s study was to show whether increasing experience in laparoscopic partial nephrectomy (group 1: less experience, and group 4: greater experience), resulted in lower IT, fewer complications, or a lower effect on surgical margins, amongst other factors. In the same way as our study, the experience referred to a single surgeon. The study was only capable of demonstrating that experience predicts the IT a statistically signi cant way (p<0.001). They also conclude that the acceptable IT having become experienced is 20 minutes, and that the number of surgeries required in order to be able to achieve this is approximately 150.
Also, as regards the planning of the surgery, the anatomical and morphometric characteristics of the tumour are crucial, as revealed by numerous studies carried out on these aspects. In fact, today there are several standard classi cations and international consensus in order to categorise kidney tumours according to their location, whether they affect the urinary tract or not, and whether the tumour is predominantly exophytic or endophytic, amongst other factors.
These are the classic score systems for renal tumours: PADUA, RENAL or the C-Index. All of these classi cation systems not only seek to unify criteria at international level with regard to their use, but also to determine the surgical di culty of a PN and to predict complications derived from this approach, with the aim of ensuring the applicability of the surgical programming of LPN.
Continuing in this line, in 2011, Zhamshid Okhunov et al (16) carried out a study, founding that any of the three score systems used (PADUA, RENAL or the C-index), independently predicted the IT (p<0.001) and the degree to which kidney function was affected (modi cation of serum creatinine pre-surgery vs. postsurgery) (p<0.001), following an LPN.
In 2015, Darren Desantis et al (17), in a study whose primary aim was to determine whether there is any association between the different scoring systems used for kidney tumours with surgical complications in PN (open and laparoscopic), concluded that out of the variables used on the different scales, tumour diameter and the endophytic and exophytic tumour portions independently predicted complications in PN (p<0.05).
In his study, Desantis took into account variables that we consider important, such as the exophytic or endophytic tumour portion, although the variable 'experience' was not analysed as such.
One interesting study published in 2018 by Sachin Yallappa et al (21), analysed both the classic PADUA and RENAL systems globally and individually for each of their items, in relation to the prediction of results following PN. They demonstrated that IT is independently associated with the PADUA score (p=0.016), the RENAL score (p=0.032), and with some of the variables included in the RENAL system, such as tumour radius (R) (p=0.004), the exophytic/endophytic component (E) (p=0.022), or the polar lines (L) (p=0.01). It also studied the relationship between the different variables included on the scales with surgical complications, demonstrating that the location of the tumour in the renal sinus (p=0.008), medial/lateral localisation (p=0.029), and affectation of the renal collecting system (p=0.006), are independently associated with renal complications.
With the aim of re ning the classic scoring systems for kidney tumours, new classi cations have appeared, which are mainly based on the classic scales (PADUA, RENAL), such as the study published in 2019 by Vicenzo Ficarra et al (22), in which the SPARE system was presented (Simpli ed PADUA RENAL), as a novel classi cation system for kidney tumours, in which the number of variables of the classic PADUA system are reduced from six to four (affectation of the renal sinus, exophytic tumour portion, tumour size, and renal border affected). The authors demonstrated that SPARE independently predicts both the overall post-surgical complications (p<0.001), as well as the major post-surgical complications (p=0.001), and concluded that their grading system could replace the classic PADUA system to evaluate the complications of tumours suitable to PN.
The study of Ficarra et al (22), was performed on a large sample of 531 patients, treated through to PN with different approach modalities (open=237, laparoscopic=152, and robotic=142). Through their analysis, they demonstrated how SPARE predicts global complications in both open (p=0.004), laparoscopic (<0.001) and robotic nephrectomy (p=0.009), or through multivariate analysis, age and SPARE also predict global complications. They also analysed each of the variables that make up SPARE, showing through univariate analysis how all of the components of SPARE predict global complications. We nd the results of this analysis interesting, as it breaks down the capacity of the new scale to predict complications in the different modalities of approach of a PN.
In 2014, Scott Leslie et al (23) de ned a new variable of great interest: the Tumor Contact Area (TCA), which they describe as the portion of the tumour in intimate contact with the renal parenchyma. This new concept has raised, per se, numerous new studies in order to determine its relation with surgical complexity, postoperative complications or its relation with loss of renal function after PN. Leslie et al. in their study showed that there was a statistically signi cant association between TCA ≥20cm² and, surgical time ≥4h (p=0.012), hospital stay (p=0.0007), and complications (p=0.037); no signi cant association with IT was found (p=0.820).
The new concept of TCA they introduced, whose calculation was made using the formula 4πr² (where r=radius of the tumour), and required software that shows renal images in 3D. The TCA could be equivalent to the morphometric variable " endophytic tumour area" that we propose in our study, and for whose calculation the use of mathematical formula or speci c complementary software was not necessary, so we think that the estimate is simpler.
In terms of results, and in contrast to the study by Leslie et al, we were not able to demonstrate any association of our variable ETA with complications, but we found a statistically signi cant association between ETA and IT.
In this same line, and in reference to TCA, in 2019, Vincenzo Ficarra et al (24), attempted to demonstrate the association of TCA both with post-surgical complications and with renal function impairment (expressed as glomerular ltrate), and showing that TCA independently predicts changes in renal function after PN (p=0.005), but without nding any statistically signi cant differences with respect to complications.
They did observe, however, that TCA (≥ 20 cm² vs <20 cm²) predicts IT (p<0.001). As for the repercussion on renal function after PN, we did not take this variable into account in our work.
TCA continued to be the subject of numerous studies. In 2016, Po-Fan (25), established a mathematical formula to more precisely calculate the TCA of renal tumours on imaging techniques (CT), proving at the same time how TCA is a better predictor of loss of renal function than the classic RENAL grading scale, COR AUC of TCA: 0,86 vs COR AUC of RENAL: 0,69. The formula devised was: TCA= 2πrd (where r is the radius of the tumour, and d its depth in the renal parenchyma). This mathematical formula proposed by Hshieh et al, to calculate TCA undoubtedly provides accuracy in its measurement, but its application is laborious, and it considers that all renal tumours are spherical, which entails a certain degree of imprecision; our way of measuring morphometric variables does not require any mathematical formula, something that many entail a certain degree of imprecision, but it is applicable to all renal tumours, regardless of their geometric shape. This feature makes it a tool that is both easy to apply and reproducible.
Finally, and based on the study presented by Po-Fan Hshieh et al, in 2018 Chalairat Suk-Ouichai et al (26), presented a study in which they showed that, using the formula TCA= 2πrd, TCA is associated with loss of renal function independently, only in the case of exophytic masses (p=0.01), but not in the case of endophytic masses (p=0.27). Neither complications nor IT were taken into account.
We believe that experience is a variable to be considered in every complex surgical procedure. Moreover, it is di cult to compare studies on complex surgical techniques if experience is not taken into account. Also, as regards the pre-surgical planning of an LPN, either using graduation scales for tumour lesions, such as the classic PADUA or RENAL, or complex mathematical formulas that require associated software to calculate the tumour contact area, these make the planning of an LPN into an additional laborious process, far from the intended purpose, which is to facilitate a fast, comfortable, simple presurgical management, and something that we consider to be of great importance: that it is reproducible at any hospital. These are precisely the objectives we intend to achieve when we propose the set of morphometric variables presented in this paper, together with the method of measuring them. This method loses precision if we compare it with the mathematical formulas presented, and which we consider to be one of the weak points of our work; however, our method is su ciently precise to predict important variables such as IT and complications.

Conclusions
In the light of the results obtained in our study, we can conclude that: ETA and surgical experience are independently associated with IT in LPN.   Figure 1 the total tumour area is 8.34 cm², line of contact 3.25 cm, and the exophytic tumour area would be 4.93 cm².