Search Results
A total of 206 articles were searched:120 from Pubmed, 24 from Ovid,20 from Cochrane Library and 42 from Web of science. After carefully inspection of these articles, 8 studies including 2385 patients published between 2015 and 2018 were finally enrolled in our meta-analysis. The detail processes of study selection were showed in the flow diagram (Fig. 1). Among them, five studies were from Japan, one study was performed in USA, China and Republic of Korea, respectively. OR and 95%CI were extracted directly or calculated indirectly in 8 studies. In our meta analysis, not every association has been evaluated using every paper included in the meta-analysis; four studies revealed the association between radiological tumor size༞2 cm and STAS, three studies illuminated the relation between pure solid nodule and STAS, five studies depicted the connection between part-solid nodule and STAS, two studies demonstrated the association between percentage of solid component(PSC)༞50% and STAS. The PSC was calculated as follows: [maximum diameter of the solid component/maximum diameter of the lesion] × 100, where the maximum diameter of the lesion includes both ground-glass opacity and solid component in CT image. The characteristics of the enrolled studies were shown in Table 1.
Table 1
The basic characteristics of enrolled studies.
author | year | region | N(M/F) | year | STAS(+/-) | subtype | stage | Lobectomy/sublobar Resection | NOS | CT-based features |
Kadota et al[4] | 2015 | Japan | 411(164/247) | 68 | 155/256 | ADC | I | 291/120 | 7 | part-solid nodule. |
Shiono et al[15] | 2016 | Japan | 318(149/169) | 70 | 47/271 | ADC | I | 202/116 | 5 | pure solid nodule. |
Dai et al[10] | 2017 | China | 383(178/205) | 60 | 116/267 | ADC | I | 364/19 | 6 | radiological tumor size༞2 cm. |
Masai et al[16] | 2017 | Japan | 508(248/260) | 66 | 76/432 | ADC | I | 0/508 | 7 | part-solid nodule. |
Toyokawa et al[17] | 2018 | Japan | 327(153/174) | 69 | 191/136 | ADC | I-IV | 235/84 | 7 | radiological tumor size༞2 cm,pure solid nodule,part-solid nodule. |
de Margerie-Mellon et al[18] | 2018 | USA | 80(27/53) | 68 | 40/40 | ADC | NA | NA | 6 | radiological tumor size༞2 cm, part-solid nodule. |
Kim et al[19] | 2018 | Republic of Korea | 276(129/147) | 59 | 92/184 | ADC | I-III | 226/50 | 6 | pure solid nodule, part-solid nodule. |
Toyokawa et al[20] | 2018 | Japan | 82(40/42) | 71 | 31/51 | ADC | I | 0/82 | 7 | radiological tumor size༞2.0 cm. |
The association between radiological tumor size༞2 cm and STAS in ADC.
The results of the association are showed in Fig. 2. Four studies presented the data to evaluated the association between radiological tumor size༞2 cm and STAS in our meta-analysis. Considering the heterogeneity (I2 = 62.8%, P = 0.045).Therefore, a random-effect model was applied. Our results clearly depicted that there is no significant relationship between radiological tumor size༞2 cm and STAS with the combined OR of 1.47(95% CI:0.86–2.51; Fig. 2). The CT-based feature of radiologic tumor size larger than 2 cm cannot be used as a biomarker to predict STAS. This finding shows that tumors larger than 2 cm in CT image are not necessarily more likely to spread through air spaces in ADC.
The association between pure solid nodule and STAS in ADC .
In our mate-analysis, Three included studies evaluating the relationship between pure solid nodule in CT image and STAS were analyzed. Since there is no heterogeneity (I2 = 0.0%, P = 0.858),Fixed-effect model was applied. Our result revealed a strong association between pure solid nodule and STAS in ADC with pooled OR of 3.10(95%CI2.17-4.43,Fig. 3). Therefore, combined effect provided the evidence that pure solid nodule in CT image were more likely to spread through air spaces, which might be a marker to predict STAS in ADC.
The association between part-solid nodule and STAS in ADC.
For the relationship between part-solid nodule in CT image and STAS in ADC. The five studies enrolled into our mate-analysis were analyzed by random effect model because of larger heterogeneity (I2 = 72.3%,p = 0.006,Fig. 4).From our mate-analysis, our result clearly showed that there is a significant association between part-solid nodule and STAS in ADC. Therefore, The part-solid nodule in CT scan may be more likely to appear STAS in ADC(combined OR:3.10,95%CI:2.17–4.43,Fig. 4) .
The association between percentage of solid component( PSC)༞50% and STAS in ADC.
Percentage of the solid component (PSC) was defined as the ratio of maximum diameter of the solid component to the tumor average diameter multiplied by 100% in CT image. Two studies presented data to evaluated the association between PSC༞50% and STAS in our mate-analysis. Fixed effect model was applied because of little heterogeneity (I2 = 42.6%,p = 0.187,Fig. 5).PSC༞50% was a significant independent predictor in the diagnosis of STAS in ADC from our result with combined OR of 2.95(95%CI:1.88–4.63, Fig. 5).de Margerie-Mellon et al[18] reported that high proportion of solid component diameter to tumor average diameter as CT manifestation in pulmonary adenocarcinomas was presented as a predictive biomarker in the diagnosis of STAS. From our mate-analysis, the tumor with PSC ༞50% is more likely to spread through air spaces in ADC.
The outcomes were similar whether fixed-effects models or random-effects models were used. Publication bias was evaluated by Begg’s funnel plot and Egger’s linear regression test, Egger’s test ( p > 0.05) showed that there was no significant publication bias of studies included this meta-analysis. The shapes of the funnel plot are symmetric visually (Fig. 6) and no proof of publication bias was obtained.