Detection rate of VPI between single-block and dual-block groups
The destruction of elastic layers of pleura was questionable on the block of the H&E slides from the same tumor in some cases. There was controversial about which one of blocks with pleural involvement or not (Fig. 1A, 1B and 2A, 2B). With the application of dual blocks for elastic stain, tumor cells invasion beyond elastic layers was visualized and the status of pleural invasion was clearly assessed (Fig. 1C, 1D and 2C, 2D). Generally, the detection rate of VPI was 9.1% (474/5182), 17.7% (573/3235) in single-block and dual-block group, respectively, which showed significantly difference between two groups (P < 0.001). In dual-block group, 410 cases showed one tissue block with VPI positivity and 163 cases with two-blocks VPI positivity. Comparing with single-block group, the VPI-positive rate increased by 8.6% by dual-block elastic stain as shown Table 1.
Table 1
Visceral pleural involvement confirmed by single and dual-block ES in peripheral 8419 NSCLC patients
Variables | VPI (%) | Non-VPI (%) | Total | P value |
Single-block ES | 474 (9.1) | 4710 (90.9) | 5184 | |
Dual-block ES | 573 (17.7) | 2662 (82.3) | 3235 | ༜0.001* |
one block positive | 410 (12.7) | | | |
Two blocks positive | 163 (5.0) | | | |
VPI, visceral pleural invasion; ES, elastic stain. |
*VPI positivity between Single-block ES group and dual-block ES group. |
The role of dual-block elastic stain in evaluating VPI status for staging of NSCLC patients with Lymph Node Negative and Tumor size ≤ 4 cm.
Cases of 6008 (T1, tumor size ≤ 3cm) and 804 (T2a tumor size, ༞3cm to ≤ 4cm ) NSCLC patients without lymph node metastasis were retrospectively reviewed. For comparing the different VPI positivity in 6008 NSCLC patients (T ≤ 3cm), the incidence of VPI were 6.3% (235/3730) in single-block group and 12.0% (273/2278) in dual-block group, respectively. The application of dual-block elastic stain had significant correlation with higher detection rate of pleural involvement than that in single-block elastic stain (P༜0.001). In single-block group, 235 of 3730 (T ≤ 3cm, stage IA) cases with VPI positivity were upstaged to T2a (stage IB). Compared with single-block group, 5.7% patients with tumor size ≤ 3cm (stage IA) were additionally upstaged to T2a (stage IB) in dual-block group. In VPI-positivity patients of dual-block group, 202 cases (74.0%) were VPI positive in one block and 71 cases (26.0%) were VPI positive in both two blocks (Table 2). However, similar incidence of VPI were not found between single-block group (16.8%, 82/488) and dual-block group (17.1%, 54/316) for elastic stain in NSCLC patients with pT2a (P = 0.916) (Table 3). In addition, the VPI was significantly more frequent (16.8% versus 8.5%, P༜0.001) in T2a (> 3 to ≤ 4 cm) tumors than that in T1 (≤ 3 cm) tumors (Table 4).
Table 2
VPI positivity between single-block and dual-block group of 6008 NSCLC patients with tumor size ≤ 3cm.
Variables | VPI (%) | Non-VPI (%) | Total | P value |
Single-block ES | 235 (6.3) | 3495 (93.7) | 3730 | |
Dual-block ES | 273 (12.0) | 2005 (88.0) | 2278 | ༜0.001* |
one block positive | 202 (8.9) | | | |
two blocks positive | 71 (3.1) | | | |
VPI, visceral pleural invasion; ES, elastic stain. |
*VPI positivity between Single-block ES group and dual-block ES group. |
Table 3
VPI positivity between single-block and dual-block group of 804 NSCLC patients with tumor size(> 3 to ≤ 4cm)
Variables | VPI (%) | Non-VPI (%) | Total | P value |
Single-block ES Dual-block ES one block positive two blocks positive | 82 (16.8) 54 (17.1) 34 (10.8) 20 (6.3) | 406 (83.2) 262 (82.9) | 488 316 | 0.916* |
VPI, visceral pleural invasion; ES, elastic stain. |
*VPI positivity between Single-block ES group and dual-block ES group. |
Table 4
VPI positivity difference between T1(≤ 3cm) and T2a (༞3cm to ≤ 4cm) NSCLC patients.
Variables | VPI (%) | Non-VPI (%) | Total | P value |
T1 (≤ 3cm) T2a (༞3cm to ≤ 4cm) | 508 (8.5%) 136 (16.9%) | 5500 (91.5%) 668 (83.1%) | 6008 804 | ༜0.001 |
Correlation between clinicopathological features and visceral pleural involvement in peripheral NSCLC Patients
The clinical and pathological characteristics of NSCLC patients are listed in Table 5. There were 4514 female patients and 3905 male patients, with a median age of 64 years (range: 25–89 years). Overall, VPI occurred in 1047 of 8419 cases (12.4%). The percentage of male patients with VPI was higher than that of female patients (P༜0.001). However, the presence of VPI had no difference in terms of age (P = 0.364). Furthermore, VPI positivity was associated with tumor size larger than 3 cm (P༜0.001), lymphovascular invasion (P༜0.001), lymph nodes involvement (P༜0.001), and the presence of STAS status also played a significant role in the incidence of VPI (P༜0.001).
Table 5
Correlation between VPI and clinical-pathologic characteristics of 8419 peripheral NSCLC patients
Characteristics | Patients(n) | VPI | | Non-VPI | P value |
| number | % | | number | % |
Total | 8419 | 1047 | 12.4 | | 7372 | 87.6 | |
Sex | | | | | | | |
| Female | 4514 | 491 | 10.9 | | 4023 | 89.1 | |
| Male | 3905 | 556 | 14.2 | | 3349 | 85.8 | < 0.001 |
Age (year) | | | | | | | |
| < 60 | 2439 | 291 | 11.9 | | 2148 | 88.1 | |
| ≥ 60 | 5980 | 756 | 12.6 | | 5224 | 87.4 | 0.364 |
Tumor size(cm) | | | | | | | |
| ≤ 3cm | 6667 | 718 | 10.8 | | 5949 | 89.2 | |
| > 3cm | 1752 | 329 | 18.8 | | 1423 | 81.2 | < 0.001 |
Lymph nodes | | | | | | | |
| Negative | 6932 | 646 | 9.3 | | 6286 | 90.7 | |
| Positive | 1212 | 360 | 29.7 | | 852 | 70.3 | < 0.001 |
| NA | 275 | 26 | 9.5 | | 249 | 90.5 | |
STAS status | | | | | | | |
| Negative | 6788 | 652 | 9.6 | | 6136 | 90.4 | |
| Positive | 1631 | 396 | 24.3 | | 1235 | 75.7 | < 0.001 |
Lymphovascular invasion | | | | | | |
| Negative | 7974 | 897 | 11.2 | | 7077 | 88.8 | |
| Positive | 445 | 150 | 33.7 | | 295 | 66.3 | < 0.001 |
Pathologic subtype* | | | | | | |
Invasive adenocarcinoma** | 7481 | 901 | 12.0 | | 6580 | 88.0 | |
| non-mucinous adenocarcinoma | 7270 | 901 | 12.4 | | 6369 | 87.6 | |
| mucinous adenocarcinoma | 210 | 0 | 0.0 | | 210 | 100 | < 0.001 |
squamous cell carcinoma | 672 | 78 | 11.6 | | 594 | 88.4 | 0.725 |
Others*** | 211 | 54 | 25.6 | | 157 | 74.4 | < 0.001 |
VPI, visceral pleural invasion. NA, no answer (patients did not undergo lungs and hilar lymph nodes dissection). STAS, tumor spread through air space; |
visceral pleural invasion. NA, not available (patients did not undergo systemic hilar and mediastinal lymph nodes dissection). STAS, tumor spread through air space; |
*26 lymphoepithelioma-like carcinoma, 23 mixed carcinoma and 7 cases with two different primary tumors were excluded; |
** include 210 invasive mucinous adenocarcinoma /collioid carcinoma, 3 enteric adenocarcinoma and 1 low-grade fetal adenocarcinoma; |
***compared with invasive adenocarcinoma; |
****include 67 large cell carcinoma, 78 sarcomatoid carcinoma, 65 adenosquamous carcinoma, and 1 poorly differentiated carcinoma with SMARC4-deficient. |
In addition, the incidence of VPI in patients with poor differentiated histologic subtypes, including 67 cases of large cell carcinoma, 78 cases of sarcomatoid carcinoma, 65 cases of adenosquamous carcinoma, and 1 poorly differentiated carcinoma with SMARCA4-deficient, was 25.6% (54/210), which was much higher than that in adenocarcinoma (12.0%, 901/7481) and squamous cell carcinoma patients (11.6%, 78/672), respectively (P༜0.001). However, the percentage of VPI had no difference between adenocarcinoma and squamous cell carcinoma (P = 0.725). Remarkably, none of invasive mucinous adenocarcinoma cases (0/210) present pleural involvement, irrespective of tumor size in our study (Fig. 3).
Correlation between growth pattern of peripheral invasive adenocarcinoma and VPI status
As shown in Fig. 4, The ROC analysis of the histogram parameters between VPI presence and VPI absence showed that AUC of well, moderate and poor differentiation were 0.263 (95% confidence interval [CI]:0.246–0.280), 0.544 (95% CI: 0.518–0.570), and 0.720 (95% CI:0.698–0.743), respectively, highlighting that lepidic histology was a protective factor of VPI, while micropapillary and solid histology were risk factors of VPI, and the combination of micropapillary and solid growth pattern was the most beneficial parameter for predicting the extent of VPI (P༜0.001). Micropapillary and solid predominant histology were significant factors for predicting pleural involvement (AUC = 0.720, cut-off value of 3.5 × 10 − 3 mm2/sec, sensitivity 74.5%, specificity 69.6%) (as shown in Table 6).
Table 6
Univariate analysis of correlation between different growth pattern of adenocarcinoma and the presence of VPI
histological comparison | AUC | 95% confidence interval | P value |
lower-bound | upper-bound |
Lepidic | 0.263 | 0.246 | 0.28 | < 0.0001 |
acinar and papillary | 0.544 | 0.518 | 0.57 | < 0.0001 |
micropapillary and solid | 0.720 | 0.698 | 0.743 | < 0.0001 |