Rectal cancer (RC) is among the most prevalent gastrointestinal cancers [1], characterized by significant tumor heterogeneity [2]. Neoadjuvant chemoradiotherapy followed by total mesorectal excision is a standard treatment for locally advanced rectal cancer [3]. Although the locoregional recurrence rate has decreased due to this treatment, distant metachronous metastases still occur in over 20-30% of cases [4,5]. The varying biological characteristics of rectal cancer may result in different treatment responses and prognoses [6]. Patients who are diagnosed with the same tumor-node-metastasis (TNM) stage of RC frequently exhibit vastly different clinical outcomes, despite receiving the same therapeutic interventions [7]. While the TNM staging system remains the predominantly used framework for directing treatment choices and forecasting prognosis, it is crucial to take into account additional pathological elements, such as tumor budding, lymphovascular invasion (LVI), perineural invasion, and extramural venous invasion [8-10]. A recent investigation has highlighted that LVI, rather than the traditionally prioritized extent of tumor invasion, serves as a significant risk factor for metastasis in early-stage RC [11]. Patients exhibiting LVI+ demonstrate a considerably worse prognosis and overall survival when compared to those with LVI -. Identifying LVI before surgery is crucial for informing treatment strategies in RC [12]. LVI offers important information regarding the tumor's biological behavior and its likelihood of metastasizing, thus improving the precision of disease staging. This information enables clinicians to tailor treatment approaches, which helps prevent overtreatment in patients with LVI -. In cases where LVI is present, it may be essential to adopt more aggressive treatment methods, such as postoperative chemotherapy, radiotherapy, or more extensive surgical resections, to reduce the chances of recurrence and metastasis. In summary, LVI serves as a crucial prognostic factor, providing essential insights into the tumor's potential for metastasis and recurrence, which has significant clinical implications for predicting postoperative outcomes, including recurrence, metastasis, and disease-free survival. Therefore, a preoperative non-invasive assessment of LVI, particularly in patients with LVI +, may be instrumental in guiding clinical practice.

Conventional magnetic resonance imaging (MRI) is particularly effective in observing large blood vessels and has moderate sensitivity and high specificity to distinguish LVI [13]. However, it can noninvasively and accurately assess extramural vascular invasion with a diameter greater than 3 mm. In contrast, it is challenging to evaluate vascular invasion with a diameter less than 3 mm, as well as intramural vascular and lymphatic infiltration, due to the limitations in resolution [14,15]. In addition, pathological biopsy is the gold standard for preoperative diagnosis. However, this invasive approach may not reflect the LVI status of the entire tumor due to possible sample errors [16]. Therefore, effectively assessing LVI in RC via preoperative imaging remains a significant challenge.

Intravoxel incoherent motion (IVIM) is a functional imaging method that can display the diffuse movement of water molecules and blood flow, microcirculation perfusion, thereby improving the diagnostic accuracy of RC [17,18]. The IVIM model has the potential to identify perfusion effects using multiple b-value samples and biexponential curve fit analysis [19], which can lead to three parameters, namely the true diffusion coefficient (D), the pseudo-diffusion coefficient (D*), and the perfusion fraction (f). Li et al [20]. discovered that the D value obtained from IVIM of LVI+ patients was significantly higher than that of LVI- patients with RC, but only 42 patients were included in this study, making it more difficult to accurately assess the feasibility of IVIM. In another study, Kan et al [21]. revealed that the D value was helpful to assess the prognostic factor related to RC, while the AUC of the D value for evaluating LVI was only 0.646. Currently, studies utilizing conventional IVIM quantitative parameters have demonstrated a relatively low diagnostic efficiency in detecting LVI, with AUC values ranging from 0.629 to 0.646 [22,23].

Tumor heterogeneity has emerged as a significant biomarker for tumor staging and prognosis [24]. Sub-region analysis, an innovative imaging post-processing technique, has demonstrated its potential in exploring intratumor heterogeneity in recent years. Prior research has underscored the importance of sub-region analysis derived from either single-modality or multiparametric MRI in the investigation of intratumor heterogeneity in RC [25-27]. However, to our knowledge, the application of sub-region analysis based on IVIM-MRI to evaluate LVI status and prognosis in patients with RC has not been reported. Thus, this study aimed to assess the fe

Rectal cancer (RC) is among the most prevalent gastrointestinal cancers [1], characterized by significant tumor heterogeneity [2]. Neoadjuvant chemoradiotherapy followed by total mesorectal excision is a standard treatment for locally advanced rectal cancer [3]. Although the locoregional recurrence rate has decreased due to this treatment, distant metachronous metastases still occur in over 20-30% of cases [4,5]. The varying biological characteristics of rectal cancer may result in different treatment responses and prognoses [6]. Patients who are diagnosed with the same tumor-node-metastasis (TNM) stage of RC frequently exhibit vastly different clinical outcomes, despite receiving the same therapeutic interventions [7]. While the TNM staging system remains the predominantly used framework for directing treatment choices and forecasting prognosis, it is crucial to take into account additional pathological elements, such as tumor budding, lymphovascular invasion (LVI), perineural invasion, and extramural venous invasion [8-10]. A recent investigation has highlighted that LVI, rather than the traditionally prioritized extent of tumor invasion, serves as a significant risk factor for metastasis in early-stage RC [11]. Patients exhibiting LVI+ demonstrate a considerably worse prognosis and overall survival when compared to those with LVI -. Identifying LVI before surgery is crucial for informing treatment strategies in RC [12]. LVI offers important information regarding the tumor's biological behavior and its likelihood of metastasizing, thus improving the precision of disease staging. This information enables clinicians to tailor treatment approaches, which helps prevent overtreatment in patients with LVI -. In cases where LVI is present, it may be essential to adopt more aggressive treatment methods, such as postoperative chemotherapy, radiotherapy, or more extensive surgical resections, to reduce the chances of recurrence and metastasis. In summary, LVI serves as a crucial prognostic factor, providing essential insights into the tumor's potential for metastasis and recurrence, which has significant clinical implications for predicting postoperative outcomes, including recurrence, metastasis, and disease-free survival. Therefore, a preoperative non-invasive assessment of LVI, particularly in patients with LVI +, may be instrumental in guiding clinical practice.

Conventional magnetic resonance imaging (MRI) is particularly effective in observing large blood vessels and has moderate sensitivity and high specificity to distinguish LVI [13]. However, it can noninvasively and accurately assess extramural vascular invasion with a diameter greater than 3 mm. In contrast, it is challenging to evaluate vascular invasion with a diameter less than 3 mm, as well as intramural vascular and lymphatic infiltration, due to the limitations in resolution [14,15]. In addition, pathological biopsy is the gold standard for preoperative diagnosis. However, this invasive approach may not reflect the LVI status of the entire tumor due to possible sample errors [16]. Therefore, effectively assessing LVI in RC via preoperative imaging remains a significant challenge.

Intravoxel incoherent motion (IVIM) is a functional imaging method that can display the diffuse movement of water molecules and blood flow, microcirculation perfusion, thereby improving the diagnostic accuracy of RC [17,18]. The IVIM model has the potential to identify perfusion effects using multiple b-value samples and biexponential curve fit analysis [19], which can lead to three parameters, namely the true diffusion coefficient (D), the pseudo-diffusion coefficient (D*), and the perfusion fraction (f). Li et al [20]. discovered that the D value obtained from IVIM of LVI+ patients was significantly higher than that of LVI- patients with RC, but only 42 patients were included in this study, making it more difficult to accurately assess the feasibility of IVIM. In another study, Kan et al [21]. revealed that the D value was helpful to assess the prognostic factor related to RC, while the AUC of the D value for evaluating LVI was only 0.646. Currently, studies utilizing conventional IVIM quantitative parameters have demonstrated a relatively low diagnostic efficiency in detecting LVI, with AUC values ranging from 0.629 to 0.646 [22,23].

Tumor heterogeneity has emerged as a significant biomarker for tumor staging and prognosis [24]. Sub-region analysis, an innovative imaging post-processing technique, has demonstrated its potential in exploring intratumor heterogeneity in recent years. Prior research has underscored the importance of sub-region analysis derived from either single-modality or multiparametric MRI in the investigation of intratumor heterogeneity in RC [25-27]. However, to our knowledge, the application of sub-region analysis based on IVIM-MRI to evaluate LVI status and prognosis in patients with RC has not been reported. Thus, this study aimed to assess the feasibility of utilizing sub-region analysis based on IVIM MRI to predict LVI status and prognosis in patients with RC.

Inclusion Criteria:

• (1) patients underwent preoperative abdominal pelvic IVIM-MRI examination; (2) patients underwent surgical resection within 2 weeks after IVIM-MRI examination; (3) complete postoperative pathological data, including LVI status.

Exclusion Criteria:

• (1) radiotherapy or chemotherapy before surgery; (2) preoperative metastatic disease; (3) poor image quality; (4) incomplete clinical data.

400 patients diagnosed with rectal cancer from the sichuan cancer hospital were consecutively recruited in this study