PHG is an important sequalae of portal hypertension and can lead to chronic blood loss with resultant refractory anemia and even liver transplantation. Blood flow congestion resulting from portal hypertension by different mechanisms is considered the main cause of PHG. This leads to imbalances between mucosal defense and injury factors caused by mucosal hemodynamic alterations thus inducing PHG (9).
Higher prevalence of PHG is found in patients with severe PH, advanced liver disease, and after esophageal varices eradication. On the other side, PHG patients with cirrhosis showed low incidence of acute GI bleeding; incidences ranged from 2.5 to 30%, with the greatest occurrences being reported in patients with severe PHG (10).
PHG is diagnosed mainly by upper GI endoscopy as it is seen in gastric mucosa as characteristic mosaic-like pattern with or without red spots. The pathogenesis of PHG was proven to involve venous congestion with gastric mucosal capillary dilation (11).
I-SCAN is a new technology based on software filter that digitally enhances high-definition endoscopic images and renders color tone, sharpness, and contrast. Three algorithms of I-SCAN can be used to get the desired the level of enhancement: Surface Enhancement enhances light-to-dark contrast for easier demarcation of edges and flat lesions, Contrast Enhancement enhances areas of low intensity for better identification of depressed lesion and Tone Enhancement for improved mucosal structure assessment by increasing the illumination and emphasis on vascular features (12). These algorithms are combined to give three I-SCAN settings used as follows: I-SCAN 1 to detect the lesions where only SE is applied to refine the subtle surface abnormalities without altering the brightness. I-SCAN 2 mode is used for characterization of lesions by combining SE and TE to enhance both minute mucosal changes and vessel structures. I-SCAN 3 adds CE to the endoscopic image with SE and TE and it helps in demarcation of lesions by digitally adding blue color to its darker edges (13).
The aim of this study was to assess the privilege of using I-SCAN over conventional white light endoscopy in the diagnosis of Portal Hypertensive Gastropathy among liver cirrhosis patients
All I-SCAN modes (1, 2 &3) showed statistically significant difference against WLE in detection of mosaic pattern in fundus of the stomach (P < 0.001) with both I-SCAN 1 &2 having the highest true positive rate (95.3%). This gives us the conclusion that I-SCAN 2 is the best mode to detect mosaic pattern compared to WLE and other I-SCAN modes with sensitivity 95.3%, specificity 51.9% and accuracy 87.7%.But; Achim et al., (2016); results come against our conclusion as they reported that I-SCAN 2 got lower accuracy than I-SCAN 1 in diagnosing mosaic pattern.
Cross tabulation of WLE red spots findings versus different I-SCAN modes findings shows that I-SCAN 2 is the only I-SCAN mode to have statistically significant difference against WLE. Although, I-SCAN 3 had the highest true positive rate (75.6%) versus (69.3%) for I-SCAN 2. Both I-SCAN 2 &3 shared sensitivity value 100% but regarding the specificity I-SCAN 2 got 52.8% against 39.6% for I-SCAN 3. Regarding the red spots examination, I-SCAN 3 detected more positive cases with red spots of PHG (112, 72.1%) than WLE and other I-SCAN modes and regarding the severity both I-SCAN 2 &3 detected the same number of cases with severe red spots (26, 16.9%).
The results of (14) agree with our results pointing to the ability of I-SCAN 2 to better detect red spots of PHG but our study was expanded to the use of I-SCAN 3 which was able to detect more cases of severe PHG by numbers as proved by the high true positive rate despite not achieving statistically significant difference against WLE as in the case of I-SCAN2.
Other studies were interested in the role and efficacy of virtual chromoendoscopic modalities in the diagnosis of PHG like (15) that used NBI with magnification to correlate and explain the WLE PHG findings with the observed microcirculation changes. His results concluded that NBI has higher accuracy than WLE in detection of PHG with lower cut off values of different non-invasive predictors of PHG. Also (16) used both NBI and I-SCAN with magnification to correlate the microcirculatory findings with the histopathological findings and concluded that I-SCAN detected the same number of cases with mild PHG but more cases with severe PHG than WLE. In this study NBI detected the same number of cases with severe PHG but milder PHG cases than WLE &ISCAN. This gives NBI the advance over I-SCAN in detection of mild PHG hence early detection and treating the condition before proceeding to the complications related to its severity.
A study of microvascular architecture of PHG related to the presence of red spots with intramucosal hemorrhage found that Low platelet count and the resulting bleeding tendency may contribute to intramucosal hemorrhage in the pathogenesis of PHG leading to the morphological appearance of red spots as an endoscopic finding (11).
The median value of platelets count in the included patients in our study was 93 (×103/µL) ranging from 27 up to 211.
Another study reported the findings of PHG in 448 out of 611 patients (73.3%) with mean value of platelets count 107 ± 66(17). While Gjeorgjievski and Cappell demonstrated the inverse correlation between the severity of PHG and platelets count as patients with mild PHG had platelets count mean value 132 ± 100.7 and count with severe PHG = 102.8 ± 68.8.(18) Also ;Mandhwani and his colleauges in their study mentioned that the mean platelets count in the studied population was 113.91 ± 69.02(19).
The comparison between the platelet counts in patients with mosaic pattern versus those with red spots seen by standard WLE examination denotes that red spots are related to lower platelet count hence more severe disease and elevated PH.
I-SCAN modes related to higher cut off values of platelet count in patients with present red spots indicating better ability and sensitivity of I-SCAN modes to detect red spots earlier. I-SCAN 2 has the highest mean value of platelet count cut off (100.93 ± 52.26).
Regarding mosaic pattern and platelet count we found that both I-SCAN 1 &2 share the highest platelet count cut off mean value (99.82 ± 48.48) by a small margin of superiority over WLE (99.75 ± 49.63).
As regard portal vein diameter, associations between PV diameter measured by abdominal ultrasonography and PHG criteria found by all used endoscopic techniques (WLE, I-SCAN 1, 2 &3) didn’t show any significant correlation.
Normal PV diameter is less than 10 mm, with a greater than 20–30% increase with food and respiration. In portal hypertension, the PV is dilated (> 13 mm), with absent or less than 20% variation with respiration as reported by Elbarbary and his colleauges(20). The patients enrolled in their study achieved highly significant increase in the PV diameter compared with the controls with mean value 14.2 ± 1.6 mm in the patient group as an indicator of PH.
I-SCAN 3 detection of red spots is associated with smaller diameter of PV (13.37 ± 2.61 mm) representing the ability of this mode to detect red spots in earlier stages of the disease compared to the other techniques used in examination.
Mosaic pattern association with PV diameter didn’t show any significant difference with all endoscopic techniques as the smallest diameter (13.59 ± 2.46 mm) with WLE and the largest diameter (13.64 ± 2.41 mm) with I-SCAN 3 with 0.5 mm difference in the mean value.
The presence of red spots in standard WLE examination is related to larger splenic diameter as detected by abdominal ultrasonography representing more elevation of portal pressure and more severity of the disease despite not achieving a statistically significant correlation. In contrast this doesn’t apply to association with the presence of mosaic pattern.
Kim and his colleagues reported statistically significant association between the severity of PHG and splenic diameter as patients with severe PHG had splenic diameter of 13.1 ± 2.4 cm while the splenic diameter with mild PHG was 12.2 ± 2.5 cm, and splenic diameter in cases without PHG was 10.7 ± 2.9 cm (21). While Mandhwani and his colleauges reported the splenic diameter among their studied population with mean value 14.65 ± 3.18 cm (19).
Splenic size shows significant correlation (p < 0.05) with red spots detection using I-SCAN 2 (p = 0.011) and I-SCAN 3 (p = 0.047). While detecting red spots both modes (I-SCAN 2 &3) are related to lower cut off mean values of spleen diameter (15.47 ± 4.43 cm, 15.7 ± 4.57 cm respectively) compared to WLE (16.91 ± 2.25 cm) representing its ability to identify red spots in earlier stages of the disease that correlates with less severe PH.
The higher values of splenic diameter found in our patients compared to Kim and his colleauges’ study could be justified as the Egyptian population have their own parameters regarding the ultrasonographic organometery which are higher than those recorded by other studies as reported by El Sharkawy and his colleauges (22).
As regards the different scoring systems applied to the population of our study and its correlation to the PHG criteria found by different endoscopic techniques; we found that Child Pugh score got a statistically significant association only with mosaic pattern detected by WLE & me-SCAN3. APRI score achieved statistically significant correlation with mosaic pattern detected with WLE & I-SCAN 1. Also, with both mosaic pattern and red spots detected with I-SCAN 2 &3. Regarding Fib-4 score, mosaic pattern detected with I-SCAN 3 was the only category that achieved statistically significant association. MELD score couldn’t achieve any significant correlation with any of the findings by all the endoscopic techniques.