| Abstract|| |
Background: Perineural invasion (PNI) is correlated with adverse survival in several malignancies, but its significance cervical cancer remains to be clearly defined. The objective of this study was to determine the association between PNI status and clinical outcomes in clinically localized surgically treated cervical cancers. Materials and Methods: We reviewed clinical records and pathology slides of 111 patients with cervical cancer treated with surgery at a single academic center. PNI was evaluated for presence, number of foci per slide, involved largest nerve size, and topographically (intratumoral vs. extratumoral). Association with these parameters, clinicopathologic characteristics and survival were analyzed. Results: The analysis demonstrated that PNI in cervical cancer was significantly correlated with parametrial invasion, tumor size, resection margin involvement, lymphovascular invasion, lymph node (LN) metastasis, depth of stromal invasion, necrosis, and higher stage disease (P < 0.005). Similarly, PNI density and mean size of the nerve involved were also associated with advanced stage (P < 0.005). In the multivariate analysis, PNI was not an independent prognostic factor for disease-free and overall survival. However, in the advanced stage cases and LN (+) cases, PNI is significantly associated with lower overall survival (43 vs. 58 months and 36 vs. 60 months, respectively, P < 0.005). Conclusions: The presence of PNI is accompanied by high-risk factors for recurrence. Overall survival rate is significantly reduced in PNI (+) patients. Although PNI itself is not an independent prognostic factor, PNI has a significant prognostic impact on overall survival in patients with advanced stage and/or Node (+) cervical cancer.
Keywords: Advanced stage, perineural invasion, prognosis, uterine cervical cancer
|How to cite this article:|
Vural C, Bayrak BY, Muezzınoglu B, Yucesoy I. Perineural invasion is a valuable prognostic factor in advanced stage and/or Node (+) cervical cancer. Indian J Pathol Microbiol 2017;60:27-32
|How to cite this URL:|
Vural C, Bayrak BY, Muezzınoglu B, Yucesoy I. Perineural invasion is a valuable prognostic factor in advanced stage and/or Node (+) cervical cancer. Indian J Pathol Microbiol [serial online] 2017 [cited 2019 Dec 14];60:27-32. Available from: http://www.ijpmonline.org/text.asp?2017/60/1/27/200021
| Introduction|| |
The best indicator for growth of a malignant tumor is the spread of the tumor locally, as well as for metastatic deposits, destruction of the extracellular matrix and the cells of the tumor gaining the ability to move away from the primary tumor., While vascular and lymphatic canals are well-known routes for metastatic spread, nerves are a less conspicuous route for spreading.,,, Spreading through nerves is called perineural invasion (PNI) and defined as the presence of tumoral cell infiltration in or around a nerve.,, PNI, which is frequently detected in pancreas, biliary tract, head-neck, and prostate tumors, inhibits local control of the tumor because neoplastic cells can progress along the nerve tracts. PNI is also a risk factor for recurrence., In addition, it is considered to be associated with aggressive course and poor survival rate.,,,,,,, The College of American Pathologists recommends notification of PNI status in prostate and pancreas carcinomas because of potential prognostic implication. Additionally, PNI is a necessary component in pathological evaluations, as the presence of PNI substantially influences strategy and decisions concerning adjuvant therapy in head-neck cancers. However, our knowledge of the prevalence and prognostic significance of PNI in cervical cancers is limited, and there is no consensus among authorities on this issue.,,
Cervical cancers Stages I–IIA are considered as an early stage disease and are conventionally treated by radical histerectomy + bilateral salpingo-oophorectomy + bilateral pelvic para-aortic lymphadenectomy., Lymph node (LN) metastasis, parametrial invasion, and resection margin involvement are well-known high-risk factors that enhance recurrence rates in postoperative cervical cancer patients and serve as indicators for adjuvant therapy for the prevention of loco-regional recurrence. The prognostic significance of intermediate risk factors such as tumor size, deep stromal invasion and lymphovascular invasion (LVI) is debatable, and they are considered important for the recurrence of disease.,, However, because of postoperative recurrence in approximately 20%–50% of the patients without LN metastasis or resection margin involvement, research is ongoing to determine whether additional risk factors should trigger postoperative adjuvant chemotherapy.,,
The present study aimed to investigate the prevalence and prognostic significance of PNI, as well as its relation with other predictive and prognostic factors, in surgically treated cervical cancers.
| Materials and Methods|| |
The present study comprised a total of 111 cervical cancer cases that were surgically treated between 2006 and 2014 and the surgical materials of which were examined by Kocaeli University, Faculty of Medicine, Department of Pathology. Of the patients, 104 underwent radical hysterectomy and pelvic and/or para-aortic lymphadenectomy, while seven only underwent conization. The present study included squamous cell carcinoma (SCC), adenosquamous carcinoma, and adenocarcinoma of the cervix, but rare types of cancer such as small cell and neuroendocrine carcinoma were excluded. Accordingly, the cases were examined in two groups: SCC and adenocarcinoma + adenosquamous carcinoma.
Pathological findings and clinical courses of the patients were reviewed by examining electronic patient files.
Recurrence, which was defined by the presence of radiological findings on computed tomography or magnetic resonance imaging, as well as localization and date of recurrence, patients' status, latest follow-up visits, deaths due to the disease, and clinical stage at diagnosis were recorded. Staging was done in accordance with International Federation of Gynecology and Obstetrics guidelines.
For the purpose of evaluating cervical cancers, the entire cervix in the resection material was sampled except for five cases. All slides for these samples were re-evaluated by two pathologists, who were blind of the clinical course of the patient, in terms of the presence of PNI, density of PNI, diameter of the greatest nerve involved, localization of the nerve involved according to the tumor, and percentage of perineural area involved by the tumor. Tumor characteristics such as tumor size, histological subtype, LVI, parametrial invasion, depth of stromal invasion, lymphocytic response, necrosis, status of resection margins, and LN metastasis were recorded.
PNI was defined as the presence of cancer cells either in, around or through the local peripheral nerve fibers regardless of the extent of tumoral involvement in perineural tissue in the slides stained with hematoxylin and eosin [Figure 1].,,, Any detectable PNI was stated as PNI (+), regardless of whether one or multiple nerves were affected in the individual case. Cases with PNI were coded as PNI (+) (n = 34) and cases without PNI were coded as PNI (–) (n = 77).
|Figure 1: Perineural invasion; the presence of cancer cells either in or close to and surrounding the local peripheral nerve fibers|
Click here to view
All PNI foci, if any, in all slides were evaluated. Density of PNI was calculated for each case using the following formula: Number of PNI foci/number of sections with tumor. The diameter of the largest nerve fiber infiltrated was measured and recorded. The cases were dichotomized according to the percentage of tumoral infiltration (<50% or ≥50%) around the nerve involved. Moreover, when considering the localization of each of identified PNI focus according to the tumor, localizations in the tumor were classified as intratumoral and those out of the tumor were classified as extratumoral.
Parametrial spread was defined as tumoral infiltration of the stroma and/or perineural area in the parametria.
For the depth of stromal invasion, the cases were evaluated in two groups as <2/3 and ≥2/3 according to the tumoral infiltration of the internal 2/3 of the cervical wall or the external 1/3.
Metastasis to the parametrial, pelvic or paraaortic LN, or to both was considered LN metastasis.
Based on previous studies, evaluation of peritumoral inflammatory response, i.e., lymphocytic response was done semiquantitatively in one microscopic field using an objective with 10-fold magnification (Olympus BX-50 field diameter 2.2 mm). Accordingly, the cases were evaluated in three groups as none-mild lymphocytic response, moderate lymphocytic response and severe lymphocytic response.
For resection margins, parametrial and/or vaginal resection margin involvement was considered positive.
All statistical analysis was performed using the Statistical Package for the Social Sciences version 20.0 (Armonk, NY, USA: IBM Corp). The Chi-square test and Fisher's exact test were used to assess the relationship between PNI status and clinicopathologic characteristics. Survival was calculated by the Kaplan–Meier method and compared using the long-rank test. Overall survival was defined as the time from the beginning of primary therapy to death due to any cause. For multivariate analysis, the independent prognostic impact of patient- and tumor-related factors on survival was investigated using the Cox proportional hazards model. Statistical significance was defined at P < 0.05.
| Results|| |
PNI was detected in 30.6% (34/111 cases) of the cases. The relationship between PNI status and clinicopathological characteristics is summarized in [Table 1].
|Table 1: Relation between presence of perineural invasion and clinicopathological characteristics|
Click here to view
The mean age of the patients was 48.95 ± 12.20 years. While the mean age of PNI (+) cases was 47.69 ± 11.25 years, the mean age of PNI (–) cases was 51.79 ± 13.88 years (P = 0.102).
PNI was associated with high risk factors such as LN metastasis (P = 0.001), resection margin involvement (P = 0.002) and parametrial invasion (P = 0.001), as well as intermediate risk factors such as tumor size (P = 0.002), depth of stromal invasion (P ≤ 0.001) and LVI (P = 0.001). Moreover, while there was a relationship between PNI and necrosis (P = 0.001), no relationship was observed with histological subtype (P = 0.814) or lymphocytic response (P = 0.109).
With regard to LN metastasis, the mean number of metastatic LNs was 3.65 ± 7.80 and was 5.16 ± 10.23 in the PNI (+) group and 1.73 ± 0.28 in the PNI (–) group. While the number of metastatic LNs was higher in PNI (+) patients, it was not statistically significant (P = 0.417).
There was a significant correlation between PNI and tumor stage. Of the patients with PNI, 82.4% were in the advanced stage (P < 0.001).
It was observed that the presence of PNI (6 vs. 28 cases), density of PNI (0.062 vs. 0.227), mean size of the nerve involved (86 vs. 158 µm), tumor size (<4 vs. >4 cm), LVI, necrosis, depth of stromal invasion (<2/3 vs. >2/3) and lymphocytic response were associated with advanced stage disease (P < 0.05). There was no correlation between the clinical stage and both the PNI location and the percentage of tumoral infiltration around the nerve involved (P > 0.05).
Follow-up data, which were lacking for only 11 patients, were available for 90.1% of the patients. In this retrospective analysis, the mean follow-up period for patients overall was 31.47 ± 23.65 months (range 1–94 months). Kaplan–Meier curves for disease-free survival rate (DFS) and overall survival rate (OS) were henceforth constructed and compared using the long-rank test. Comparing 5-year DFS between PNI (+) and PNI (–) patients, no significant difference was observed (48.2% [95% confidence interval [CI]: 34.8%–61.7%] vs. 41.2% [95% CI: 34.2%–48.2%], P = 0.329) [Figure 2]. Five-year OS was substantially lower for PNI (+) patients (80.9% [95% CI: 70.1%–91.6%] vs. 95.3% [95% CI: 91.4%–99.3%]) [Figure 3] and this was statistically significant (result of Long-rank test: Chi-square = 5.254, P = 0.022).
|Figure 2: Kaplan–Meier curve for disease-free survival in patients with perineural invasion and those without perineural invasion|
Click here to view
|Figure 3: Kaplan–Meier curve for overall survival in patients with perineural invasion and those without perineural invasion|
Click here to view
The cases were grouped according to stage (early stage and advanced stage) and the status of LN metastasis (Node [–] and Node [+]), and PNI (+) and PNI (–) groups were evaluated in terms of survival [Table 2]. There was no statistically significant difference between these two groups among early stage and Node (–) patients, whereas it was observed that survival times were substantially shortened among advanced stage (P = 0.047) and Node (+) (P = 0.039) patients in the PNI (+) group. In Kaplan–Meier analysis, advanced stage PNI (+) patients had significantly low OS as compared to advanced stage without PNI (42.6% [95% CI: 27.0%–58.2%] vs. 58.0% [95% CI: 44.7%–71.2%], P = 0.047). These patients did not differ in DFS (38.3% [95% CI: 25.0%–51.7%] vs. 51.4% [95% CI: 36.7%–66.2%], P = 0.147). When patients with LN metastasis and PNI (+) were analyzed both OS (35.5% [95% CI: 24.3%–46.7%] vs. 60.1% [95% CI: 46.4%–73.9%], P = 0.039) and DFS (32.8% [95% CI: 23.2%–42.3%] vs. 60.1% [95% CI: 43.3%–77.0%], P = 0.015) were significantly low.
|Table 2: Mean overall survival time for patients with cervical cancer with and without perineural invasion|
Click here to view
Multivariate Cox regression analysis included LVI, parametrial invasion, histological subtype, resection margin involvement, age, tumor size (≤4 and >4 cm), necrosis, depth of stromal invasion, lymphocytic response, status of LN involvement and PNI. Within this framework, it was observed that PNI is not an independent prognostic factor. While parametrial invasion, resection margin involvement, age, LN metastasis, and tumor size were independent prognostic factors (P < 0.05); LVI, histological subtype, necrosis, depth of stromal invasion, and lymphocytic response were not independent prognostic factors.
| Discussion|| |
Whereas it has generally been accepted that cervical cancers spread either directly or throughout the vascular/lymphatic canal, there has been a limited number of studies on the role of PNI in the spread of these tumors. In the literature, conflicting results have been reported on the prevalence and prognostic importance of PNI in cervix cancers.,, This is thought to have resulted from limited case numbers in some studies or differences in study design. Another important reason is the lack of a concise, universally accepted definition for PNI;,,, there are numerous different definitions in the literature.,,,,, Dunn et al. defined the minimum histopathological criteria required to make the diagnosis of PNI as follows: “In the presence of a malignant condition, PNI may be diagnosed based on detection of cytologically malignant cells in the perineural space of nerves. In suspicious cases, observing total or almost total circumferential involvement is supportive, as is the presence of perineural tracking in tangential sections and intraneural involvement.”
In the literature, the prevalence of PNI was reported between 7% and 41.7% in different studies investigating PNI in uterine cervical cancers.,,,, In this retrospective analysis that consists of 111 cervical cancer cases, the prevalence of PNI was found to be 30.6%, which is consistent with the range reported in the literature. It was observed that PNI was associated with high-risk factors such as LN metastasis, resection margin involvement and parametrial invasion, as well as intermediate risk factors such as tumor size, depth of stromal invasion and LVI, and also with advanced stage and necrosis (P < 0.05), but not with histological subtype (P = 0.814) or lymphocytic response (P = 0.109). Horn et al. conducted a study in 2009 including 194 cases with Stages IBI–IIB cervical cancers and similar to the findings of the present study, reported substantially higher PNI risk in patients with pelvic LN metastasis, deep stromal invasion, and parametrial invasion. The same authors conducted a study in large-size cervix carcinomas with occult parametrial invasion, and recommended precise and sometimes numerous sampling to detect PNI in large, local and advanced stage tumors, and stated that it was important to obtain a clear resection margin particularly in Stage IIB when PNI was detected in radical hysterectomy specimens. They also stated that they embedded these tissues into different blocks while evaluating parametrial resection margin in daily practice. All cervical and parametrial tissues were sampled in the majority of our specimens. Contrary to the finding in our study, Ozan et al. stated that there is no significant difference between PNI (+) and (–) groups in terms of parametrial invasion, as well as in terms of histological subtype, tumor stage, mean tumor size, millimeter of stromal invasion, pelvic LN metastasis, resection margin involvement, recurrence, and mortality rate, and that PNI is related closely only to deep cervical stromal invasion, vaginal-uterine involvement, and LVI. A few years later, Meinel et al., who determined a significant correlation between the presence of PNI and parametrial invasion and deep cervical stromal invasion (41% in those <66% and 16.9% in those >66%, P = 0.001), demonstrated decreased PNI in uterine cervix carcinomas with significant lymphocytic response (18.8%–58.8%, P < 0.001), which is different from the findings of the present study. Elsahwi et al. stated that PNI is associated with large tumor size, parametrial invasion, LVI and uterine spread, but not with LN metastasis, depth of invasion, tumor stage or spread to the vaginal margin. Moreover, these researchers reported that the mean age was 45.9 years in PNI (–) cases, whereas it was 53 in PNI (+) cases and that they found statistically significant the difference between age and PNI (P = 0.01). In the present study, there was no significant difference between PNI (–) and PNI (+) cases in terms of mean age (51.79 vs. 47.69, P = 0.102). In their 185-case study comprising early-stage cervix cancer cases, Cho et al. also could not find a significant relationship between PNI and mean age (P = 0.980), and reported that PNI showed a significant relationship with LN metastasis, resection margin involvement, parametrial invasion, depth of invasion and LVI (P < 0.05). Moreover, they reported that the mean number of metastatic LN was 4.4 in PNI (+) cases and 1.5 in PNI (–) cases and that there is a significant relationship between the number of metastatic LNs and PNI (P = 0.026). In the present study as well, the mean number of metastatic LN was higher in the PNI (+) group (1.73 ± 0.285 vs. 5.16 ± 10.23); however, this was not statistically significant (P = 0.417).
Consistent with information in the literature, the present study found a significant correlation between PNI and tumor stage.,,, It was notable that tumors with PNI, higher PNI density and larger mean size of the nerve involved, as well as the larger tumors, tumors with LVI, necrosis and depth of stromal invasion were advanced stage tumors (P < 0.05).
Beitler et al. carried out a retrospective analysis in 1997 and evaluated the role of intraoperative interstitial brachiotherapy during total pelvic exenteration in 26 patients with recurrent cervix carcinoma. The analysis found that resection margin, LVI, PNI and combinations of these unfavorably influenced disease-specific survival and local tumor control. Horn et al. reported that the 5-year OS significantly decreased in PNI (+) cases (51.1% vs. 75.6%, P = 0.001), whereas the 5-year recurrence-free survival rate decrease in PNI (+) cases (78.3% vs. 69.5%) had no statistical significance. In the analysis, which was performed by dichotomizing the cases as Node (–) and Node (+), they reported that this prognostic effect continued in Node (+) patients, but was absent in Node (–) patients, and that the prognostic effect of PNI in OS showed “borderline significance” in these groups. Ozan et al. reported that patients with versus patients without PNI have shorter OS, whereas Cho et al. reported that the presence of PNI has no effect on OS or DFS. In the present study, 5-year OS decreased in PNI (+) patients versus PNI (–) patients (80.9% vs. 91.6%), and this was statistically significant (P = 0.022). There was no significant difference between 5-year DFSs of PNI (+) and (–) patients (48.2% vs. 41.2%, P = 0.329). When the cases were analyzed by being dichotomized as Node (–)/Node (+) according to early stage/advanced stage and LN metastasis, it was determined that the survival rate significantly decreased among advanced stage (P = 0.047) and Node (+) (P = 0.039) patients in the PNI (+) group (life times; 43 vs. 58 months and 36 vs. 60 months, respectively, P < 0.05). This was not previously reported in cervical cancer. Although the case number is limited, reporting PNI in advanced stage and or LN (+) can guide choice of adjuvant therapy and surveillance following treatment of cervical cancer.
Multivariate analysis in previous studies revealed that PNI is an independent prognostic factor.,, Memarzadeh et al. reported that multivariate analyses revealed PNI, LVI, large tumor size (>4 cm) and depth of invasion as significant predictive factors for recurrence (P < 0.05) and that histopathological findings, particularly in parametrium, may be used as a factor in selecting patients for adjuvant therapy. Like the present study, publications that report the limited significance of PNI as an independent predictive factor also emphasize that PNI is associated with a combination of high- and intermediate-risk factors, all of which are considered when making decisions regarding the need for adjuvant therapy., Tavares et al. conducted a study in 2009 to investigate prognostic factors in 301 cases with FIGO I and II uterine cervix cancer, and stated that the most significant prognostic factors are LN metastasis, tumor size, resection margin involvement and necrosis (P < 0.05) and that PNI is not associated with prognosis. In the present study, multivariate Cox regression analysis revealed that PNI is not an independent prognostic factor. While parametrial invasion, resection margin involvement, age, LN metastasis and tumor size were identified as independent prognostic factors, it was observed that LVI, histological subtype, necrosis, stromal depth and lymphocytic response were not independent prognostic factors.
The present study found the prevalence of PNI as 30.6% in surgically treated cervical cancers, and it was observed that overall survival significantly decreased in such patients. It was determined that PNI is not an independent prognostic factor for these tumors, but is associated with well-known high and intermediate risk factors such as parametrial invasion, LN metastasis, resection margin involvement, tumor size, depth of invasion, and LVI. Although the predictive value of PNI for the recurrence of cervical cancers is generally limited, we recommend reporting perineural space involvement since it may help with clinical decision making and surveillance in advanced stage and/or LN (+) cases thus allowing patients at a greater risk of decreased survival to be offered more intensive therapy. However, further studies are needed to support these findings and to assess the exact prognostic value of PNI in cervical cancers.
| Conclusion|| |
We recommend reporting perineural space involvement in cervical cancer, although predicitve value for reccurence is limited, In patients with advanced stage and/or LN (+) presence of perineural invasion necessisate more intensive theraphy.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Horn LC, Meinel A, Fischer U, Bilek K, Hentschel B. Perineural invasion in carcinoma of the cervix uteri – Prognostic impact. J Cancer Res Clin Oncol 2010;136:1557-62.
Meinel A, Fischer U, Bilek K, Hentschel B, Horn LC. Morphological parameters associated with perineural invasion (PNI) in carcinoma of the cervix uteri. Int J Surg Pathol 2011;19:159-63.
Cho HC, Kim H, Cho HY, Kim K, No JH, Kim YB. Prognostic significance of perineural invasion in cervical cancer. Int J Gynecol Pathol 2013;32:228-33.
Jardim JF, Francisco AL, Gondak R, Damascena A, Kowalski LP. Prognostic impact of perineural invasion and lymphovascular invasion in advanced stage oral squamous cell carcinoma. Int J Oral Maxillofac Surg 2015;44:23-8.
Elsahwi KS, Barber E, Illuzzi J, Buza N, Ratner E, Silasi DA, et al.
The significance of perineural invasion in early-stage cervical cancer. Gynecol Oncol 2011;123:561-4.
Dunn M, Morgan MB, Beer TW. Perineural invasion: Identification, significance, and a standardized definition. Dermatol Surg 2009;35:214-21.
Duraker N, Sisman S, Can G. The significance of perineural invasion as a prognostic factor in patients with gastric carcinoma. Surg Today 2003;33:95-100.
Liebig C, Ayala G, Wilks JA, Berger DH, Albo D. Perineural invasion in cancer: A review of the literature. Cancer 2009;115:3379-91.
Ozan H, Ozuysal S, Ediz B. Perineural invasion in early-stage cervical carcinoma. Eur J Gynaecol Oncol 2009;30:379-83.
Memarzadeh S, Natarajan S, Dandade DP, Ostrzega N, Saber PA, Busuttil A, et al.
Lymphovascular and perineural invasion in the parametria: A prognostic factor for early-stage cervical cancer. Obstet Gynecol 2003;102:612-9.
Singh N, Arif S. Histopathologic parameters of prognosis in cervical cancer – A review. Int J Gynecol Cancer 2004;14:741-50.
Pecorelli S, Zigliani L, Odicino F. Revised FIGO staging for carcinoma of the cervix. Int J Gynaecol Obstet 2009;105:107-8.
Hassanein AM, Proper SA, Depcik-Smith ND, Flowers FP. Peritumoral fibrosis in basal cell and squamous cell carcinoma mimicking perineural invasion: Potential pitfall in Mohs micrographic surgery. Dermatol Surg 2005;31(9 Pt 1):1101-6.
Miller ME, Palla B, Chen Q, Elashoff DA, Abemayor E, St. John MA, et al.
A novel classification system for perineural invasion in noncutaneous head and neck squamous cell carcinoma: Histologic subcategories and patient outcomes. Am J Otolaryngol 2012;33:212-5.
Ronaghy A, Yaar R, Goldberg LJ, Mahalingam M, Bhawan J. Perineural involvement: What does it mean? Am J Dermatopathol 2010;32:469-76.
Tavares MB, Sousa RB, Oliveira e Silva T, Moreira LA, Silva LT, Tavares CB, et al.
Prevalence of prognostic factors for cancer of the uterine cervix after radical hysterectomy. Sao Paulo Med J 2009;127:145-9.
Horn LC, Einenkel J, Höckel M, Kölbl H, Kommoss F, Lax SF, et al.
Pathoanatomical preparation and reporting for dysplasias and cancers of the cervix uteri: Cervical biopsy, conization, radical hysterectomy and exenteration. Pathologe 2007;28:249-60.
Beitler JJ, Anderson PS, Wadler S, Runowicz CD, Hayes MK, Fields AL, et al.
Pelvic exenteration for cervix cancer: Would additional intraoperative interstitial brachytherapy improve survival? Int J Radiat Oncol Biol Phys 1997;38:143-8.
Prof. Bahar Muezzınoglu
Department of Pathology, Kocaeli University Faculty of Medicine, Kocaeli
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2]