Determination of the positional relationship of the second Houston valve and peritoneal reflection using computed tomographic colonography and magnetic resonance imaging

The second Houston valve is used as a surrogate for estimating the position of the peritoneal reflection; however, the concordance between the positions of the valve and peritoneal reflection has not been investigated. This study aimed to clarify this positional relationship. The second Houston valve and peritoneal reflection positions were assessed using tomographic colonography and magnetic resonance imaging. In total, 117 patients were enrolled in this study. The positions of the second Houston valve and peritoneal reflection were nearly concordant, although the space between them ranged from − 20.7 to 33.9 mm. A peritoneal reflection located further from the anal verge than the second Houston valve was defined as a shallow peritoneal reflection. Male sex, high body weight, and a high body mass index were significantly correlated with a shallower peritoneal reflection, as determined by a univariate analysis (sex: P = 0.0138, weight: P = 0.0097, body mass index: P = 0.0311). A multivariate analysis revealed a significantly shallower peritoneal reflection in males than in females (odds ratio: 2.75, 95% confidence interval: 1.15–6.56, P = 0.023). The second Houston valve located near the peritoneal reflection can be a useful surrogate marker for estimating its position. In relatively heavy males, the peritoneal reflection is located more cranially than the second Houston valve.


Introduction
Rectal cancer has a high local recurrence rate (22%) which can be reduced by preoperative radiation therapy [1,2]. According to the European Society for Medical Oncology (ESMO) guidelines, preoperative chemoradiation (CRT) is indicated for lower rectal cancer but is not recommended for upper rectal cancer (occurring > 12 cm from the anal verge), which should be treated as colon cancer [3]. A meta-analysis by Cillian et al. reported a significantly lower local recurrence rate in the upper rectum than in the middle and lower rectums (odds ratio [OR], 0.495) [4].
The ESMO guidelines define the lower rectum using an endoscope to measure the distance from the anal verge [3]. However, the Japanese guidelines define the lower rectum based on the peritoneal reflection (PR) and also recommend preoperative CRT for lower, but not upper, rectal cancer [5]. The reason for this is that if the lesion is located below the PR, it drains into the internal iliac system, which often leads to pelvic lymph node recurrence and metastasis to the lung, thus requiring a treatment method different from that for other colorectal cancers. To our knowledge, no report exists on whether the prognosis changes when lower rectal cancer is defined by the anal verge or PR. However, it is necessary to accurately evaluate the position of the PR in relation to the cancerous lesion for treatment guidance.
A previous study reported difficulty in determining the position of the PR using endoscopic measurements and observed significant individual variation in its position, ranging from 8.5 to 20 cm above the anal verge [6]. Another author reported the use of magnetic resonance imaging (MRI) to identify the PR based on recognition of its membrane structure [7]. It has been observed that in 14% of cases, PR was not visible on MRI, and even when visible, the probability of diagnosing a localized lesion in the upper rectum was only 89% [8].
In Japan, the location of the second Houston valve (HV) is often considered to be concordant with that of the PR and is widely used for determining the location of the PR. The line connecting the superior border of the pubis and the inferior border of the fourth sacral vertebra is also used as a reference marker. However, in clinical practice, we occasionally find lesions preoperatively identified as being cranial to the second HV to actually be positioned below the PR intraoperatively. To our knowledge, no study has verified whether the location of the second HV is concordant with that of the PR.
We therefore assessed the positional relationship between the second HV and PR; MRI was used to identify the PR, and computed tomographic colonography (CTC) was used to identify the second HV. We also developed an accurate method of determining the position of a lesion in relation to the PR in preoperative evaluations for rectal cancer.

Patients
A total of 147 patients with rectal cancer were enrolled in this study. Patients underwent CTC and MRI between January 2013 and March 2020 before treatment at the University of Tokyo Hospital. None of the patients had previously undergone pelvic surgery. The study protocol was approved by the Ethics Committee of the University of Tokyo [No. 3252- (10)].

Procedures
CTC was performed after total colonoscopy with full bowel preparation using polyethylene glycol electrolyte lavage solution (Niflec ® ; Ajinomoto Pharma, Tokyo, Japan) with an auto-injector (PROTO CO2L; ACIST Medical Systems, Tokyo, Japan). An enema tube was inserted through the anus with the patient in the prone position, and CO 2 was automatically injected at a pressure of 20 mmHg. Threedimensional computed tomography of the rectum was performed using an imaging workstation (Ziostation2 ® ; Ziosoft Inc., Tokyo, Japan). MRI was performed using a 1.5-T MRI scanner (MAGNETOM Avanto, Siemens Healthineers) with the administration of an intramuscular spasmolytic. Sagittal T2-weighted images were then obtained.

CTC and MRI evaluations
To fuse the information from CTC and MRI, a reference line was developed based on the bone information from both images. The reference line connected the superior border of the pubis and the inferior border of the fourth sacral vertebra. In CTC, the distance from the anterior wall of the second HV to the reference line, and in MRI, the distance from the anterior wall of the PR to the reference line was measured in the lateral position (Fig. 1). To evaluate the positional relationship between the second HV and PR, the distance between the second HV from PR was measured by subtracting the height of the PR from the height of the second HV. We defined a shallow PR as one in which the PR was more cranial than the second HV. A deep PR was defined as one in which the PR was more caudal than the second HV.

Statistical analyses
Statistical analyses were performed using the JMP software program (version 15.0; SAS Institute, Inc., Cary, NC, USA). The relationship between anatomical indices (sex, age, height, weight, and body mass index [BMI]) and the second HV from PR was evaluated using Spearman's correlation coefficient. The heights of the second HV, PR, and second HV from PR were analyzed by a univariate logistic regression analysis using anatomical indices (sex, age, height, and weight), and a multivariate analysis was performed by a logistic regression analysis using indices with P values < 0.10. Statistical significance was set at P values < 0.05.

Depth of second Houston valve, peritoneal reflection, and distribution of the second HV from PR
PR using MRI, and second HV using CTC were assessed in 147 patients. Among these, PR was not detected in 26 patients (detection rate 83%), and in 4 patients, it was difficult to assess second HV due to insufficient air inflation (detection rate, 97%). As a result, 117 patients were enrolled in this study. Patient variables are shown in Table 1.
We examined the positional relationship among the second HV, PR, and the reference line. The second HV was cranial to the reference line in the majority of cases (89%), as shown in Fig. 2a. The median value was 8.2 mm, and the measurement range was 15.9-24.6 mm. In contrast, the position of the PR was nearly concordant with that of the reference line, as shown in Fig. 2b. The median value was 2.4 mm, and the measurement range was 18.9-26.5 mm. The positional relationship between the second HV and PR was examined. The PR was slightly deeper than the second HV, as shown in Fig. 2c. The median value was 4.5 mm. We found significant variation in positional relationships. In some cases, the PR was located much deeper than the second HV (max: 33 mm), as shown in the two-dimensional (2D) image in Fig. 2d. In other cases, the PR was concordant with the second HV, as shown in Fig. 2e, or it was located much shallower than the valve (minimum: -20 mm), as shown in the 2D image in Fig. 2f.

Correlation between second HV from PR and patient variables
The correlation of second HV from PR with patient variables is shown in Fig. 3. The PR was significantly shallower in males than in females (0.0138). No correlation was found between the age and height (age: 0.6787, r = − 0.0387; height: 0.1231, r = − 0.1433). A higher body weight and BMI values were correlated with a shallower PR (weight: 0.0097, r = − 0.2382, BMI: 0.0311, r = − 0.1994). We performed a logistic regression analysis to examine the relationship between the location of the second HV and PR from the reference line and the interrelationship between the location of the second HV and PR with patient variables. The PR was located significantly more cranially than the reference line in patients with a higher body weight and BMI (OR: 0.96, 95% confidence interval [CI]: 0.93-0.99, P = 0.0426). It was also significantly shallower in males than in females (OR: 2.75, 95% CI: 1.15-6.56, 0.023).

Discussion
To our knowledge, this is the first study to examine the positional relationship between the second HV and PR using CTC and MRI.
The position of the PR was nearly concordant with the reference line (Fig. 2b). Yamamoto et al. showed that the PR is nearly concordant with the line connecting the superior border of the pubis and the inferior border of the fifth sacral vertebra, based on their study using eight cadavers [9]. In addition, they considered the position of the PR in cadavers to be approximately 1 cm caudal to the living body due to the relaxation of the anorectal muscles, which is generally consistent with the results of this study. As an indicator for estimating PR, the reference line is considered useful.
The positions of the second HV and PR were nearly concordant, with a displacement of up to 3 cm on the cephalocaudal side (Fig. 2c). Two earlier studies reported the distance from the anal verge to the PR; one measured the distance from the anal verge to the PR intraoperatively using an endoscope with a median of 13.2 cm and a range of 8.5-21.0 cm [6], and the other used an endoscope to obtain a median distance of 8.8 cm and a range of 6.4-11.2 cm [10]. The distance from the anal verge to the second HV was also measured in 400 patients using a rigid endoscope, with a mean of 9.4 cm and a range of 7-12 cm [11]. Thus, from the anal verge, the measured distances between the valve and PR were generally similar, with variations within a certain range, but there were large individual differences. We found no report that evaluated the positional relationship between the second HV and PR in the same patient. In the present study, the second HV tended to be located more cranially than the PR. We examined the correlation between the position of the second HV and PR and patient variables. A univariate analysis revealed that the second HV was caudal to the PR, and the PR was significantly shallower in males with a high body weight and BMI than in others (Fig. 3), whereas no correlation was observed with height or age. In the multivariate analysis, males were found to have a significantly shallower PR than females ( Table 2). Previous studies examined the correlation between the position of the PR and patient variables. Wasserman et al. reported that the distance from the anal verge to the PR was greater in cases with a high body weight and BMI [6] and that the average distance tended to be greater in males than in females, as also reported by Memon et al. [12] and Najarian et al. [13]. These observations are in agreement with the results of the present study.
We examined whether the shallower PR of the second HV in males with a high BMI was due to a shallower PR or a deeper second HV. Using the reference line as the standard, we examined whether the positioning of the valve and PR differed depending on patient variables. The PR was significantly more cranial than the reference line for heavier patients in the multivariate analysis. However, the position of the second HV was not affected by patient variables in the univariate analysis (Table 2). This suggests that the reason for the elevation of the second HV in males with a high body weight is due to a change in the height of the PR without a change in the position of the valve. The second HV is a structure in the intestinal tract, with its position not dependent on patient variables. A higher PR in males with a Representative images of the correlation between the second Houston valve and peritoneal reflection. Some cases showed the peritoneal reflection to be located deeper than the second Houston valve (d), at almost the same position (e), and shallower than the second Houston valve (f). The second Houston valve from the peritoneal reflection: the value after subtracting the height of peritoneal reflection from the height of the second Houston valve heavier body weight can be attributed to a higher amount of retroperitoneal fat than in others. This may result in a greater amount of fat below the PR, resulting in a higher position. In males, the seminal vesicles and prostate gland are located caudal to the PR, so the position of the PR is higher in males than in females.
In general, it is important to determine preoperatively whether a rectal cancer tumor is located cranial or caudal to the PR because cancers on the caudal side are associated with a higher local recurrence rate [14][15][16]. Preoperative CRT is recommended to reduce the local recurrence in these cases [17][18][19]. The location of the PR is generally evaluated by MRI [20,21], although a study reported that the PR was visible by MRI in only 74.4% of cases [7]. Another study reported that the PR could be recognized with certainty in only 68.2% of cases and could not be evaluated in approximately 30% of cases. The same study also revealed that when preoperative MRI was used to determine whether the lesion was located cranial or caudal to the PR and results were compared with intraoperative findings, the assessments were incorrect in 10% of cases [17]. Therefore, MRI is not always reliable. The positional relationship between the second HV and the lesion can be easily observed during a preoperative examination using endoscopy. Endoscopy of the rectums of 400 adults revealed that the second HV was visible in 92% of cases, while in 6.5% of cases only one HV was visible, and in 1.5% of cases no HV was present [11]. The present study reveals that the second HV can be confirmed by endoscopy when it is difficult to evaluate the PR by MRI, and conversely, the second HV can be evaluated by MRI when it is difficult to evaluate it endoscopically. It is important to make a comprehensive judgment based on multiple modalities, rather than a single modality, because the locations of the second HV and PR may differ depending on patient background. In Japan, upper and lower rectal cancers are customarily classified, and the indication for CRT is determined based on the location of the lesion in relation to the second HV. Whether dividing the upper and lower rectum based on the second HV or information obtained from MRI reflects a better prognosis is unclear. It was difficult to evaluate this in the present study because of the small number of cases and the single-center nature of the study. In the future, it will be necessary to determine whether the method based on the second HV or the one based on MRI reflects the prognosis more accurately, and thereby identify the best method to use to determine the indications for CRT.
Several limitations associated with the present study warrant mention. First, it was conducted at a single institution and only included Asians. Since the anatomical location of the second HV may differ based on ethnicity, further investigations are warranted in multi-center, multi-ethnic studies. Second, all patients had rectal cancer, which may have affected the original position of the second HV and PR due to changes associated with tumor growth, such as invasion of the intestinal wall and intestinal dilatation. It is desirable to conduct a similar study in patients without lesions. Third, CTC may cause the second HV to be pushed up cranially due to airflow, while with MRI, the position of the PR is stable due to the lack of air inflation. Thus, the second HV tends to be cranial to the PR. The evaluation was performed on surgical specimens, which were unaffected by air inflation; however, the surgical specimens had intestinal spasms, making it difficult to recognize the second HV. Intraoperative endoscopy or cadaveric studies might be useful for examining the positional relationship between the second HV and PR in realtime and evaluating the degree of displacement caused by air inflation.
Our study shows that the position of the second HV is concordant with that of the PR, although its position may differ depending on patient variables. We found that in males with a high BMI, the PR is located more cranially than the second HV. In conclusion, the location of the second HV is useful as a surrogate marker for estimating the positional relationship between the lesion and PR at the time of the preoperative endoscopic examination.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection, and analyses were performed by KKK, HZ, KS, KM, SE, YY, SS, YN, HS, and SI. The