Radiographic changes in retroperitoneal corridor dimensions during staged positioning for prone transpsoas lateral lumbar interbody fusion

Abstract

Background

Questions about the surgical corridor through the psoas muscle and the retroperitoneal space have persisted since the introduction of lateral lumbar interbody fusion (LLIF) in the lateral decubitus position. The prone transpsoas (PTP) approach maintains or restores lordosis and allows single-position access to the anterior and posterior columns, but the position of visceral structures during an actual PTP case has not been characterized.

Methods

Retrospective single-arm observational cohort of 51 consecutive adults undergoing LLIF-PTP by 4 fellowship-trained surgeons at 3 sites. The anterior-posterior (AP) distance from the anterior border of the quadratus lumborum (QL) to the posterior margin of the nearest nonfat visceral or retroperitoneal structure was measured at 3 stages: preoperative supine MRI (Stage 1), intraoperative prone CT after positioning but before incision (Stage 2), and intraoperative CT after retroperitoneal development and interbody placement (Stage 3). Multilevel cases were measured at the most caudal level. General linear modeling assessed change across stages, with exploratory subgroup analyses.

Results

Mean age was 62.8 years, 22 patients (43%) were male, mean BMI was 29.8 kg/m², and 45 cases (88%) were single-level. AP distance increased across stages (32.1 vs. 37.6 vs. 42.9 mm, p < .001), a mean gain of 10.8 mm. Across stages, distance differed by sex (p = .026); BMI category did not (p = .053), nor did number of incisions (p = .404). No approach-related complications occurred.

Conclusions

Prone positioning and staged retroperitoneal corridor development were associated with a stepwise increase in QL-to-viscera distance. The single-arm design and sample size preclude comparative or safety conclusions.

Introduction

Lateral lumbar interbody fusion (LLIF) surgery performed in the lateral decubitus position has seen limited adoption since its inception in the early 2000s [[1], [2], [3], [4]]. Despite advantages of the lateral interbody approach such as indirect decompression [5], minimal muscle dissection [[6], [7], [8]], and ability to insert cages with a wider footprint [9], questions regarding the characteristics of the surgical corridor through the psoas muscle and the retroperitoneal space have persisted [10,11]. In particular, visceral injuries including bowel, renal, and ureteric injuries have been reported [11].

More recently, the prone transpsoas (PTP) approach for LLIF has been developed. This approach capitalizes on the effects of prone positioning in maintaining or restoring lordosis, while allowing simultaneous access to the anterior and posterior columns of the spine without repositioning [3,12]. This approach has seen increased utilization over the past 5 years, with several studies reporting early clinical and radiographic outcomes [10,[13], [14], [15], [16]]. However, similar concerns regarding the spatial relationship between visceral structures and the surgical corridor have been raised regarding the PTP approach. Some authors argue that the surgical corridor through the retroperitoneal space is further narrowed in the prone position as opposed to the lateral decubitus position, bringing the bowel, in particular, into the line of the surgical corridor to the lateral disc [17,18].

A few studies have evaluated retroperitoneal corridor dimensions in prone versus lateral decubitus positioning with conflicting results [[17], [18], [19]]. Menezes et al. imaged 34 awake volunteers in both positions and found the distance from the posterior disc to the colon decreased from 41.1 mm in lateral decubitus to 20.3 mm prone, with the colon lying a mean 21.7 mm posterior to the anterior psoas margin prone compared with 5.5 mm in lateral decubitus, and concluded that the prone position leaves a smaller working window free of colon [18]. Amaral et al. imaged 24 patients in 3 positions and reported posterior retraction of the psoas in the prone position without a change in plexus position at L3–L4 or L4–L5 [19]. Dodo et al. compared supine and intraoperative prone CT in 94 patients positioned prone for posterior procedures and found a significant ventral shift of the kidneys at L2–L3 and the colon at L3–L4, yet still identified 29.6%–88.6% of retroperitoneal organs within a defined at-risk zone [20]. Godzik et al., in cadaveric and early clinical work, described posterior displacement of the psoas in the prone position along with greater soft-tissue accumulation between the skin and the spine, and converted one prone case to lateral decubitus because of abdominal girth [17]. These differences likely reflect variation in study design, including the use of nonsurgical volunteers, cadaveric models, or surgical patients positioned for posterior procedures rather than actual operative LLIF-PTP access [[17], [18], [19], [20]]. None of these settings measures the corridor after the retroperitoneum has been developed.

The primary objective of this study was to evaluate staged changes in the position of visceral structures in a consecutive series of patients undergoing LLIF-PTP by analyzing serial imaging preoperatively, intraoperatively after prone positioning, and again after retroperitoneal dissection to the interbody space. Secondary analyses explored whether corridor measurements across stages and final Stage 3 distance varied by BMI and patient sex.

Materials and methods

After Institutional Review Board approval, a retrospective observational cohort of consecutive patients at least 18 years of age who underwent LLIF-PTP using a procedurespecific positioning (PTP Positioner, ATEC Spine) and access system (Sigma PTP Access System, ATEC Spine) was identified at each participating institution. This was a single-arm study without a comparator group. Patients were excluded if they had a prior history of retroperitoneal or abdominal surgery. All cases were performed by 4 fellowship-trained spine surgeons. Screening covered an 8-month period and was limited to cases in which intraoperative navigation was used, as these were the cases for which intraoperative CT was available. Complications were captured from the operative, inpatient, and outpatient records through 6 months postoperatively. Standard demographic and surgical data were collected including age, sex, BMI, smoking history, the American Society of Anesthesiologists (ASA) grade, estimated blood loss, operative time, and any complication.

Eligible patients also had to have a minimum set of standard-of-care imaging available, including preoperative supine magnetic resonance imaging (MRI) (Stage 1), intraoperative prone computed tomography (CT) obtained after surgical positioning and any coronal adjustment but prior to abdominal incisional exposure (Stage 2), and intraoperative CT obtained after development of the retroperitoneal space and interbody implant placement and before any coronal adjustment was released (Stage 3).

Stage 1

(Fig. 1A): Patient is in the supine position with abdomen intact, no surgical exposure and no dissection of the retroperitoneum.

Fig. 1.

Representative photos illustrating patient positioning and procedural stages for LLIF-PTP: (A) Stage 1 = preoperative supine MRI; (B) Stage 2 = intraoperative prone positioning on a procedure-specific positioner with complete abdominal suspension and coronal plane adjustment applied prior to incision; (C) Stage 3 = intraoperative prone after abdominal incision(s), retroperitoneal development, and retractor-based access for interbody implant placement.

Stage 2

(Fig. 1B): Patient is positioned prone for the LLIF-PTP procedure using a procedure-specific positioner which allows for the abdomen to hang freely through the frame and for coronal adjustments to be made for the access corridor; prior to surgical exposure or dissection of the retroperitoneum.

Stage 3

(Fig. 1C): Patient is prone, following the abdominal incision, digital dissection of the retroperitoneal space, and retractor-based access to the disc space for the interbody implant insertion. The Stage 3 CT was obtained after retractor removal, so the Stage 3 measurement reflects the corridor once access had been developed rather than a corridor held open by instrumentation.

From each stage, the linear anterior-posterior (AP) distance was measured from the anterior border of the quadratus lumborum (QL) to the posterior margin of the most posterior nonfat visceral or retroperitoneal structure at that level (Fig. 2). The measured boundary was the nearest posterior structure that could limit corridor development (typically bowel/colon, kidney, ureter, or other retroperitoneal content); a fat plane alone was not considered the limiting structure. In multilevel cases, measurements were made at the most caudal interbody level only.

Fig. 2.

Representative axial images from a single patient at L4-L5 demonstrating the corridor measurement technique across procedural stages. The QL-to-viscera AP distance was measured on the operative approach side as the linear distance from the anterior border of the quadratus lumborum (QL) to the posterior margin of the nearest posterior nonfat peritoneal or retroperitoneal structure at the index level. A fat plane alone was not considered a limiting structure. Measurement calipers are shown with magnified insets at each stage so that the displayed distances are legible at print size. Images illustrate the measurement technique and were not selected to represent mean values. Nondiagnostic level annotations present in the source export of panel A have been masked; the masked region is remote from the measurement. (A) Stage 1, 9.09 mm; (B) Stage 2, 18.58 mm; (C) Stage 3, 27.78 mm.

Surgical technique and positioning protocol

LLIF-PTP was performed as previously described [3] with access to the retroperitoneal space created through finger dissection. All procedures included computer-assisted stereotactic navigated placement of minimally invasive posterior fixation. Navigation was utilized for posterior pedicle screw placement; the lateral interbody cage was placed under direct visualization through the retroperitoneal access corridor without navigation assistance.

All procedures were performed in the prone position using a procedure-specific positioning system that permits full abdominal suspension and controlled coronal plane adjustment. Positioning was performed to maintain a freely mobile abdomen throughout the case. The chest and hip pads were adjusted to avoid anterior abdominal compression and to allow the abdominal contents to fall and translate freely with gravity and coronal adjustment.

Pelvic fixation and pad placement followed the same protocol across all sites. The hip pads were positioned directly over the anterior superior iliac spines (ASIS) to provide pelvic support while leaving the lateral access window clear. Coronal plane adjustment was then applied through the articulating chest and hip pads to create a lateral bend on the planned access side. The degree of coronal adjustment was individualized to the target level and patient morphology. Lower extremity positioning was used to maintain maximal hip extension in the prone position; padding was placed beneath the knees to support the legs while maintaining the hips in extension.

Retroperitoneal access was achieved using blunt digital dissection, with mobilization of the peritoneal envelope and its contents away from the lateral abdominal wall and psoas. Fascial adhesions and reflections tethering the peritoneum to the muscular boundaries of the retroperitoneal space (including the transversalis fascia/abdominal wall, quadratus lumborum region, and psoas fascia) were released, allowing anterior fall of the peritoneum and expansion of the working corridor.

A single-incision or 2-incision technique was utilized based primarily on surgeon preference. The present study was not designed to compare incisional strategies, and the observed corridor measurements should not be interpreted as evidence supporting 1 technique over the other. In cases using 2 incisions, an initial small posterolateral incision was used to enter and develop the retroperitoneal space digitally and to free fascial tethering, followed by a second incision for retractor-based docking and transpsoas access. The posterolateral incision is placed at a reproducible access point as described in prior retroperitoneal anatomic descriptions and lateral-approach technique literature [21].

Statistical analysis

Continuous variables were summarized as the mean ± standard deviation (SD), and categorical variables as frequencies with percentages. General linear modeling was used to evaluate change in the AP distance through the 3 stages. Exploratory subgroup analyses evaluated AP distance across the 3 stages stratified by interbody level, BMI status, sex, laterality, and number of incisions. Separate regression analysis was used to identify factors associated with Stage 3 distance from the quadratus lumborum to the posterior-most border of any visceral structure. Statistical significance was defined as p < .05. All analyses were performed using SPSS v29.0 (IBM).

Results

Fifty-one (51) patients across 3 sites (4 surgeons) met inclusion criteria and were included in this analysis (Table 1). Mean age was 62.8 years. There were 22 males (43%) and 29 females (57%), and 6 patients (12%) were smokers. Mean ASA grade was 2.5 (0.5). Mean BMI was 29.8 kg/m² with 13 (25%) in the normal category, 14 (27%) overweight, 21 (41%) obese and 3 (6%) morbidly obese. The majority of cases (45, 88%) had a single-level fusion. An almost equal number had a left-sided approach (26, 51%) and a right-sided approach (25, 49%). Most patients had a single incision (32, 63%). Mean operative time was 128.0 minutes (45.5) and mean estimated blood loss was 61.3 mL (77.8).

Table 1.

Summary of demographic and surgical characteristics.

Age, y, mean (SD)	62.79 (15.28)
Male, N (%)	22 (43%)
Body mass index, kg/m2, mean (SD)	29.84 (5.93)
Category, N (%)
Normal	13 (25%)
Overweight	14 (27%)
Obese	21 (41%)
Morbidly obese	3 (6%)
ASA grade, N (%)
2	26 (51%)
3	24 (47%)
4	1 (2%)
Smoker, N (%)	6 (12%)
Fusion levels, N (%)
One-level
L2–L3	4 (8%)
L3–L4	8 (16%)
L4–L5	33 (65%)
Two-level
L2–L4	2 (4%)
L3–L5	4 (8%)
Laterality, N (%)
Left	26 (51%)
Right	25 (49%)
Number of incisions, N (%)
One	32 (63%)
Two	19 (37%)
Operative time, min, mean (SD)	128.0 (45.5)
EBL, mL, mean (SD)	61.3 (77.8)

There was a statistically significant increase in the AP distance from the quadratus lumborum to the posterior-most border of any visceral structure across the 3 stages (Table 2), with an overall within-subject effect of stage (p < .001). Mean distance was 32.1 mm at Stage 1, 37.6 mm at Stage 2, and 42.9 mm at Stage 3, corresponding to increments of 5.5 mm from Stage 1 to Stage 2 and 5.3 mm from Stage 2 to Stage 3, and a total increase of 10.8 mm. Across stages, AP distance differed by sex (between-subject effect, p = .026), whereas differences by BMI category (p = .053), interbody level (p = .622), laterality (p = .841), and number of incisions (p = .404) did not meet the prespecified threshold for statistical significance. As there were only 3 patients who were morbidly obese, these were grouped with the obese patients. Exploratory regression analysis (Table 3) identified higher BMI and male sex as the variables most strongly associated with Stage 3 distance. Number of incisions was not associated with AP distance across stages in the stratified comparison (p = .404) or with Stage 3 distance in the regression model (standardized β = 0.142, p = .306).

Table 2.

Mean (SD) QL-to-viscera anterior-posterior distance (mm) across procedural stages.

	Stage 1	Stage 2	Stage 3	p-value
Overall	32.1 (21.1)	37.6 (25.0)	42.9 (26.8)	<.001*
Stratified by interbody level				.622†
L2–L3	41.9 (45.2)	48.7 (50.0)	50.7 (49.4)
L3–L4	32.9 (20.8)	40.1 (19.9)	44.3 (23.6)
L4–L5	30.8 (18.1)	35.8 (23.2)	41.7 (25.4)
Stratified by body mass index				.053†
Normal	20.8 (23.5)	26.2 (25.9)	27.6 (22.3)
Overweight	28.9 (22.0)	36.9 (25.8)	42.2 (28.8)
Obese	40.0 (16.1)	43.7 (23.1)	51.6 (24.9)
Stratified by sex				.026†
Males	39.2 (20.2)	46.6 (25.6)	52.4 (28.6)
Females	26.6 (20.4)	30.6 (22.6)	35.7 (23.4)
Stratified by laterality				.841†
Left	30.9 (20.0)	36.2 (24.5)	42.6 (25.3)
Right	33.2 (22.5)	39.0 (26.0)	43.2 (28.8)
Stratified by number of incisions				.404†
One	30.3 (19.9)	35.4 (23.1)	39.9 (25.3)
Two	35.0 (23.2)	41.2 (28.2)	48.0 (29.1)

^⁎p-value for the within-subject effect of stage (Stage 1 vs. Stage 2 vs. Stage 3).

^†p-value for the between-subject effect of the stratifying variable across all 3 stages (L2–L3 vs. L3–L4 vs. L4–L5; normal vs. overweight vs. obese; males vs. females; left vs. right; one vs. two incisions).

Table 3.

Results of regression analysis of factors associated with distance between the linear anterior-posterior distance from the anterior-most border of the quadratus lumborum to the posterior-most border of any visceral structure at stage 3.

Variable	Standardized β-coefficient	p-value
BMI	0.484	<.001
Sex	0.356	.016
Number of incisions	0.142	.306
Interbody level	−0.107	.428
Laterality	−0.064	.707
Number of levels	−0.037	.778
Age at surgery	0.025	.864

Model includes 7 predictors in 51 patients, fewer than 10 observations per variable; results are exploratory.

There were no approach-related complications in this cohort.

Discussion

In this consecutive series of 51 LLIF-PTP cases, the measured QL-to-viscera AP corridor widened in stepwise fashion across the 3 study stages, from preoperative supine MRI to intraoperative prone CT before exposure and again after retroperitoneal access had been developed. The mean increase from Stage 1 to Stage 3 was 10.8 mm. These findings are descriptive measurements from a single operative workflow. They do not establish superiority over lateral decubitus LLIF or over standard prone positioning, neither of which was studied.

The value of this series is the operative setting in which the measurements were obtained. Prior reports addressing prone corridor anatomy have relied on awake volunteers, cadavers, or patients positioned prone for posterior-only procedures [18,20,22,23]. Those settings do not reproduce the combined effects of general anesthesia, full abdominal suspension, coronal adjustment, and active retroperitoneal mobilization during LLIF-PTP. Our findings suggest that conclusions drawn from nonoperative or posterior-only prone models may not fully reflect the corridor present during an actual prone transpsoas case.

The Stage 1 to Stage 2 change reflects the combined effect of general anesthesia, prone positioning, abdominal suspension, and coronal adjustment. This design does not allow the contribution of any 1 component to be separated. Even so, the data suggest that corridor enlargement begins before incision when the abdomen is allowed to translate anteriorly without external compression. In that sense, positioning is part of corridor development, not just preoperative setup.

The additional increase from Stage 2 to Stage 3 reflects the combined effect of retroperitoneal dissection, fascial release, retractor-based access, and interbody work performed before the Stage 3 scan [21]. We did not measure these components separately, so their individual contributions cannot be assigned. The important point is that the corridor seen on preoperative imaging is not the same as the corridor ultimately available to the surgeon. If the goal is to understand the corridor present during surgery, Stage 3 measurements are more relevant than baseline imaging alone. Because the Stage 3 scan was obtained after the retractor had been removed, the corridor was not being held open by instrumentation when it was measured. This also helps place cadaveric and volunteer studies in context, since those models capture only the preaccess condition.

The pattern across stages suggests that the lateral corridor in the prone position behaves as a potential space rather than a fixed one. Roughly half of the total increase occurred before incision, with positioning and abdominal suspension, and roughly half occurred after the retroperitoneum was developed. Other approaches that depend on developing a plane rather than finding one behave the same way, the anterior cervical approach described by Smith and Robinson being the most familiar [24]. These data show that the corridor measured on preoperative imaging understates the corridor available once access has been developed. They do not show whether one positioning or dissection technique yields a larger corridor than another, since the workflow was uniform across all 51 cases and no alternative was studied.

These findings are not in conflict with Menezes et al. The 2 studies measure different conditions. Menezes et al. imaged awake volunteers positioned for imaging rather than for surgery, and measured the posterior disc to colon distance with no access created [18]. The present measurements were obtained in anesthetized patients positioned for the operation itself, with the abdomen suspended and coronal adjustment applied, and the Stage 3 measurement was made after the retroperitoneum had been developed. The landmark pairs also differ, so the absolute values are not directly comparable. A narrower window in an awake volunteer lying prone in a scanner and a wider corridor in an anesthetized, operatively positioned and dissected patient are compatible observations. That distinction is what this series was designed to characterize.

Subgroup findings should be interpreted cautiously. Male sex was associated with greater Stage 3 distance in the stratified analysis (p = .026) and in the regression model (p = .016). BMI was associated with Stage 3 distance in regression when modeled as a continuous variable, but BMI category did not meet the prespecified threshold for statistical significance in Table 2 (p = .053). These findings should therefore be considered exploratory. They may reflect differences in body habitus or pelvic anatomy, but this study was not designed to test mechanism. Stage 3 distances were numerically greater at L2–L3 than at L4–L5, which is anatomically expected, although interbody level was not a significant predictor in regression. Incision strategy was determined by surgeon preference rather than by protocol, so we tested whether it tracked with the measured corridor; it did not, in either the stratified comparison (p = .404) or the regression model (p = .306). Pooling single- and 2-incision cases is therefore reasonable for this analysis, although the study was not designed to compare the 2 techniques and a difference smaller than this cohort can resolve cannot be excluded. Although absolute distances varied approximately twofold across strata, the increase from Stage 1 to Stage 3 fell within a narrower range in every subgroup, between approximately 7 and 13 mm. This pattern suggests that habitus and anatomy set the baseline corridor, whereas the increment observed across stages was similar regardless of stratum. This observation is descriptive and was not formally tested.

Limitations

This study has several limitations. It is a retrospective single-arm series performed within 1 device-specific operative workflow. There was no lateral decubitus comparator and no standard prone comparator without the procedure-specific positioner, so we cannot determine whether the observed corridor dimensions are unique to this workflow or reproducible with other setups. The radiographic endpoint is an anatomic surrogate. We did not correlate corridor dimensions with visceral injury, neurologic events, or patient-reported outcomes, and measurement reproducibility was not formally assessed. The regression model included 7 predictors in 51 patients, fewer than 10 observations per variable, so the model should be read as exploratory and its coefficients as unstable. Several subgroup standard deviations exceed their means, reflecting small numbers in some strata, and those strata should not be interpreted on their own. Stage 1 used supine MRI, whereas Stages 2 and 3 used prone CT, so modality and patient-state differences may have contributed to the observed measurements along with the positioning and access effects of interest. The cohort was 88% single-level, and only the most caudal level was measured in multilevel cases. Multilevel procedures may carry a different risk profile [25]. Finally, no access-related complications occurred in this series, but with n = 51 and visceral injury being rare, the absence of complications cannot be interpreted as evidence that the approach prevents such events.

Conclusions

Prone transpsoas positioning and staged retroperitoneal corridor development were associated with a stepwise increase in the QL-to-viscera corridor distance across all 3 procedural stages. The corridor widened from supine baseline to prone positioning and further increased following retroperitoneal dissection and retractor-based access. All measurements were conducted in patients actually undergoing the procedure as opposed to cadavers or live nonanesthetized volunteers and thus exemplify real-world conditions. No access-related complications were observed in this cohort; however, the study was not powered to assess the incidence of rare visceral injuries. The corridor increased in 2 numerically similar increments, with positioning and again with retroperitoneal development, indicating that the working space in this approach is developed rather than fixed.

Declaration of competing interests

M. Craig McMains reports consulting fees from Alphatec Spine, Inc. (Level F). David G. Schwartz reports consulting fees from Alphatec Spine, Inc. (Level F). Kelli M. Howell reports consulting fees from Alphatec Spine, Inc. (Level F) and Carlsmed, Inc. (Level C), and stock or stock options in Alphatec Spine, Inc. (Level I). Bryan Lee reports consulting fees from Alphatec Spine, Inc. (Level F). Yi Yuen Wang reports consulting fees from Alphatec Spine, Inc. (Level F). Siddhartha Srivastava declares no conflicts of interest.