Histopathology of pulmonary carcinomas revisited: Defining invasion in adenocarcinoma and how to stage multiple pulmonary carcinomas Erik Thunnissen, Hans Blaauwgeers, Federica Fillipello, Francesca Ambrosi INTRODUCTION Historically, the World Health Organisation (WHO) classification of pulmonary carcinomas has been based on the histological appearance of tumours in resection specimens [1,2]. Histology of these tumours is often heterogeneous and, as a rule, adequate sampling of the resected tumour and haematoxylin and eosin (H&E) staining of the sections will result in sufficient representation of the variation in tumour histology. In daily practice, pathological diagnosis of lung cancer is performed on small biopsy and/or cytology samples. Histological classification of lung cancer has gradually evolved in the last decades. Until the end of the last century, clinical treatment decisions only required a distinction between small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). In the present century, chemotherapy for squamous cell carcinoma (SCC) became different from that for adenocarcinoma (AC) and, furthermore, some acquired molecular changes were found to be predictive for response to specific, usually less toxic drugs. These developments required more detailed histological subtyping and molecular analysis, respectively. In less differentiated carcinomas, making the distinction between SCC and AC may be difficult on H&E-stained sections. This has resulted in a category of lung carcinoma 'not Erik Thunnissen MD PhD, Consultant Histopathologist, Department of Pathology, Amsterdam University Medical Center, VUmc, Amsterdam, the Netherlands Email: e.thunnissen@amsterdamumc.nl Hans Blaauwgeers MD PhD, Consultant Histopathologist, Department of Pathology, OLVG LAB BV, Amsterdam, the Netherlands Email: hansblaauwgeers@icloud.com Federica Filipello MD, Consultant Histopathologist, Department of Pathology, Michele and Pietro Ferrero Hospital, Verduno (CN) and Department of Pathology, San Raffaele Scientific Institute, Milan, Italy Email: federica.filipello@gmail.com Francesca Ambrosi MD, Consultant Histopathologist, Pathology Unit, Maggiore Hospital-AUSL Bologna, Bologna, Italy Email: fra.ambrosi@gmail.com Scanned with CamScanner 94 Histopathology of pulmonary carcinomas revin. otherwise specified' (NOS), the diagnosis made in approximately 30-40% of cases. Applying thyroid transcription factor 1 (TTF-1) and P63/40 immunohistochemical analysis in the NOS cases reduced the NOS category to a few percent. In 2011, a new AC classification was established to provide uniform terminology and diagnostic etiteria, as well as an overall approach to small biopsy/cytology specimens, and a strategy for management of small tissue samples for molecular and immunohistochemical studies (3). With an increasing need for optimal predictive immunohistochemical and molecular analyses, discussions on histopathological classification issues shifted into the background. Nevertheless, in order to reach a more globally uniform characterisation of pulmonary AC, several morphological aspects require consideration. The most burning of these concern criteria to assess invasion, including the role of 'collapse' of pulmonary parenchyma on invasion assessment. Another contentious issue in the histopathological evaluation of lung carcinoma is the question of when multiple lesions might be regarded as a single carcinogenic event or, alternatively; as multiple primary tumours. In the second section of this chapter, we will provide our insight into this issue. THE DEFINITION OF INVASION IN PULMONARY ADENOCARCINOMA In the 2014 WHO classification, the lesions conceptually preceding invasive AC are adenomatous hyperplasia and AC in situ (AIS), while in the AC category minimally invasive AC figures. In the definition of all of these, invasion is a key element. WHO 2014 and 2021 classifications define that AIS is as a localised, small (≤30 mm) AC with growth restricted to neoplastic cells along pre-existing alveolar structures (lepidic growth), lacking stromal, vascular, alveolar space and lacking pleural invasion or necrosis. Characteristically, tumour cells are arranged in continuous monolayers, sometimes with evidence of cell overlap or mild stratification. Papillary or micropapillary patterns are absent, although minor cellular tufting may be seen. Intra-alveolar tumour cells, either within the tumour or spread in airspaces in the surrounding parenchyma, are also absent. The text suggests that recognition of these characteristics is straightforward and without much ambiguity. In reality, the histological perception of architecture of pulmonary tissue is influenced by collapse of pulmonary parenchyma during surgical intervention and/or by pathological processes preceding or accompanying the development of a neoplasm. In assessing invasion in pulmonary ACs, pathologists do generally agree upon the ends of the spectrum (4): true stromal invasion (Figure 5.l), and on the other end of the spectrum (Figure 5.2) absence of invasion when tumour cells line thin alveolar walls surrounding large and unfolded alveolar 'air' spaces (4). In resection specimens, alveolar spaces, which in vivo were fully expanded, are reduced during surgery, for which the term iatrogenic collapse is used, and the resulting architectural changes in the pulmonary parenchyma have led to a gradual but significant variation in interpretation of invasion among pathologists |4,5). In a ring study conducted to resolve this issue, a subgroup of pathologists consistently interpreted a subset of cases as invasive, while another subgroup consistently interpreted the same cases as non-invasive, both using the critera of the WHO classification of pulmonary adenocarcinomas (WHOCOPAs) 14|. This study was based upon expert pulmonary pathologists reading selected microphotographic images, which precludes selection of different areas within a histologic section as cause for this difference. Scanned with CamScanner The definition of invasion in pulmonary adenocarcinoma 95 Figure 5.1 (a) Focus with true invasion: scattered infiltrative tumour cells (arrows) in desmoplastic stroma. (b) The single neoplastic cells are easily recognised in the corresponding cytokeratin 7 (CK7) staining. (c) Elastica van Gieson (EvG) shows the tumour cells are surrounded by collagen fibres (pink). Note the focal remnants of elastin fibres (black). As these were expert pulmonary pathologists, they all disposed of the minimally required knowledge in applying the WHOCOPAs. This clearly indicates that closer calibration of histologic criteria is needed. As stated by others: 'More reliable methods to differentiate histologic patterns may be necessary, including refinement of the definitions' (6). In this context, we will discuss the concept of collapse, the stromal response in pulmonary AC, and the effect of tangential tissue sectioning. Collapse in pulmonary pathology, the term 'collapse is used for two different phenomena [7): iatrogenic collapse, previously also called 'surgical collapse' (8), and biological collapse (9-13). Iatrogenic collapse (essentially an artefact of deflation) is due to the reduction of parenchymal air, blood, and lymph leading to atelectasis during surgery, resulting in a loss Scanned with CamScanner 96 Histopathology of pulmonary carcinomas revisited: Defining invasion... Figure 5.2 (a) Adenocarcinoma in situ (AIS) in partially collapsed lung with some less collapsed alveoli (more air). A focal area with slightly more stromata has smaller alveolar spaces (circle). Due to tangential cutting, pseudo- acinar and pseudo-multilayering structures and give the impression of invasive adenocarcinoma [haematoxylin and eosin (H&E) staining]. (b) Cytokeratin-7 (CK7) immunohistochemical staining (CK7), where the tangentially cut alveoli retain a'regular' impression of partially collapsed parenchymal morphology. (c) In the Elastica van Gieson (EvG) histochemical staining, the elastin fibres (black) are present in the stromal components, confirming they correspond to pre-existent alveolar structures. Note (i) that the fragmented appearance is the normal two-dimensional cross section of a fibre from a three- dimensional fibre network; and (i) that linear elastin (circle area) represents a surface (sheet) of increased elastin deposition in the third dimension. Likewise, a cross section of the alveolar surface is a line in the two-dimensional (H&E). of alveolar volume. In comparison with the in vivo state, in the post-surgery state the 'air' component shows the largest reduction in volume: post-surgical lung volume is estimated to be approximately a third of the volume in vivo. The reduction of alveolar volume has an important consequence, of which the significance for histological assessment of invasion was only recently recognised: deflation of an alveolus does not reduce alveolar wall surface area |14). Loss of alveolar volume induces partial infolding of alveolar walls, which modifies the pre-existing alveolar architecture. This may have important effects on the perception of the histological growth pattern of the tumour. When alveoli are lined with a monolayer of tumour cells in an AlS, deflation of air during surgery together with infolding (iatrogenic collapse) may result in seemingly papillary and acinar structures. When, on the other hand, alveoli are filled with tumour cells, e.g. with solid or cribriform growth, reduction Scanned with CamScanner The definition of invasion in pulmonary adenocarcinoma 97 of alveolar volume does not occur This has important consequences for the perception of invasion. The evolution of the adenocarcinomas with good prognosis concept in pulmonary AC started when, in 1995, Masayuki and colleagues (9) described two patterns of in situ AC: one with thin alveolar walls, type A (Figure 5.3), and one with thick alveolar walls, type B (Figure 5.4). Type B represents what is now known as tumour-related biological collapse (9-13), and is characterised by increased elastin content in the pre-existing alveolar architecture. The increase of condensed (contracted) elastin reduces alveolar size [13]. Monolayers of tumour cells will line alveolar walls [13]. Focally, epithelial cells may disappear, leaving small alveoli with thin stromal fibres. In 1999 the WHO classification defined the papillary subtype as invasive without incorporating the also initially described desmoplastic stroma. Subsequently, Yim et al [15] and Borczuk et al [16) reported a Figure 5.3 (a) Prominent collapsed adenocarcinoma in situ (AIS) with focal slightly less collapsed alveolar spaces filled with alveolar macrophages [red arrows, haematoxylin and eosin (H&E)]. On the right side, an interlobular septum, resected by the proliferation (read: no invasion). Within the lobules elongated, oval-shaped alveoli with some parallel orientation. Pseudo-stratification of tumour cells is present in part of the alveoli, what is due to tangential cutting. Together with the maximal collapse (hardly any air left this should not be misinterpreted for a solid proliferation. (b) Cytokeratin 7 (CK7) staining emphasises the regular/homogeneous architecture of the tumour with a pseudoglandular pattern. (c) Elastica van Gieson (EvG) stain shows vessels and in stroma fragmented elastin fibres (black). The latter proves the two- dimensional profile of collapsed alveolar walls. Scanned with CamScanner 98 Histopathology of pulmonary carcinomas revisited: Defining invasion... Figure 5.4 (a) Haernatoxylin and eosin (H&E) stain of partially collapsed adenocarcinoma in situ (AIS) with central increase in stroma. (b) Elastin staining of corresponding (Elastic Stain Core Kit, Special Stains Van Gieson CS, Ventana) showing that the stronal increase is due to increase of elastin (c) Cytokeratin 7 (CK7) staining with a regular structure compatible with monolayer of tumour cells (AIS). The combination of monolayer of tumour cells on alveolar wall (except the focal effect of tangential cutting) and stromal changes fit in the AIS category of Noguchi type B. Note that this is based on H&E only easily mistaken for invasive acinar adenocarcinoma. Of Note, angular glandular structures and increase in stromal thickness are not a discriminating feature in the determination of invasion (5]. subgroup of invasive ACs with mixed subtype and ‹0.6 cm invasion and an excellent prognosis (100% 5-year recurrence free survival) [16). In the 2011 AC classification, this subgroup was incorporated as 'minimally invasive, defined as tumours £3 cm, predominant lepidic pattern and ≤0.5 cm of invasion (3). Furthermore, in the original description of papillary carcinoma, the papillary architecture was important, but fibrovascular cores with complicated secondary and tertiary branches were considered as sufficient for diagnosis of invasion (17). However, the the initially also described desmoplastic stroma was not incorporated into the 1999 and subsequent WHO classifications. In hindsight, this omission contributed to a tendency to overdiagnose AIS as invasive AC. These observations can only be correctly understood with the recognition of pseudo-papillary and pseudoacinar structures in collapsed AIS, as they overlap with the patterns described for invasive acinar and papillary carcinoma in the WHO classifications of 1999, 2015 and 2021 [2,18, 19]. Unawareness of pseudopapillary and pseudoacinar structures in collapsed AIS explains to a large extent the low reproducibility Scanned with CamScanner The definition of inva of invasion in pulmonary ACs (4,5) and this underpins the need for modified criteria for the classification of pulmonary AC. In these modifications, morphology for recognising collapsed AIS is supported by a histochemical stain for elastin and an immunohistochemical stain for CK7. Elastin staining facilitates recognition of the pre-existing alveolar wall [7). In a tissue section, which is two-dimensional, fibres of the three-dimensional elastin fibre network are seen as small segments of elastin fibres. Since elastin fibres are not produced by epithelial (tumour) cells, the presence of elastin in a tumour point towards a component of a pre-existing alveolar wall, suggesting AlS. Classical papillary or acinar carcinomas do not produce elastin. Invasive AC with its firm consistency is usually accompanied by desmoplastic stroma, which disrupts the elastin framework. Therefore, the elastin staining pattern is useful in the differential diagnosis between collapsed AIS and invasive papillary and acinar AC, as is shown in Figures 5.5 and 5.6. When a papillary carcinoma is diagnosed on H&E according to WHOCOPA, it should have papillae but also many cross sections with thin fibrovascular Figure 5.5 (a) Adenocarcinoma in situ (AIS) with maximal iatrogenic collapse. There is hardly any air left. When using only the haematoxylin and eosin (H&E), the tumour could according to the WHO classification be diagnosed as invasive adenocarcinoma with a papillary or acinar pattern. (b) Cytokeratin 7 (CK7) staining with a regular pattern and parallel oriented alveoli. In the middle, slightly right is pseudo-stratification of tangential cutting (red circle). Note that this is a phenomenon present in part of an alveolar wall and not in the whole circumference. (c) Elastica van Gieson (EvG) staining where the thin alveolar walls are recognised by the fragmented elastin fibres. This excludes stroma made by and diagnosis of papillary or acinar adenocarcinoma. Scanned with CamScanner 100 Histopathology of pulmonary carcinomas revisited: Defining invasion... Figure 5.6 (a) Haematoxylin and eosin (H&E) stain of partially collapsed adenocarcinoma in situ (AIS) with focal pseudopapillary appearance. (b) Cytokeratin 7 (CK7) staining with a regular structure compatible with AIS. (c) Elastin staining of corresponding area (Elastic Stain Core Kit), with focal elastin in the stroma! cords, proving the stromal structure to be pre-existing alveolar wall and excluding a tumour made papillary structure. (d) Higher power of'C' with focal elastin in papillary-like structures. cores. When seemingly papillary structures are abundant and cross-sectioned papillae are missing, this also points towards collapsed AIS. Collapsed AIS has a regular pattern in cytokeratin 7 (CK7) staining, while in invasive AC an irregular pattern is present, as shown in Figure 5.7. The CK7 stain helps in finding areas without regular structure and to scrutinise these for foci with invasion in corresponding H&E and elastin stains, as shown in Figure 5.1. Not infrequently, architectural features highlighted by additional staining may be recognised also as subtle differences in colour and shape in the H&E stain. Therefore, additional stains, although helpful, may not always be necessary. A further issue is tumour size. In the WHOCOPA definitions, size is incorporated: AlS is defined as not larger than 3 cm. In minimally invasive AC, the depth of invasion is defined as <0.5 cm and for a lepidic tumour component as not larger than 3 cm. From a practical point of view, this gives clear handles for diagnosis. However, from a tumour biological point of view this does not make sense as a noninvasive tumour larger than 3 cm might still be regarded as AlS. In fact, tumours with AlS characteristics but up to 5 cm in size have been described (20,21). WHOCOPA guidelines require tumours up to 3 cm to be embedded in toto in paraffin. This supports thorough microscopic examination, allowing a diagnosis of AIS by excluding areas of invasion. A consequence of considering AIS beyond a size of 3 cm is that the whole of a (much) larger lesion should be embedded, to exclude possible foci of invasion. A dire consequence is then Scanned with CamScanner The definition of invasion in pulmonary adenocarcinoma 101 a Figure 5.7 (a) Invasive adenocarcinora with irregular architecture showing branched glands, strands and single tumour cells. (b) Cytokeratin 7 (CK7) staining highlights the haphazardly arranged tumour cells (irregular pattern). (c) Elastica van Gieson (EvG) staining showing that the glands are not surrounded by elastin fibres. Note on the right side some pre-existing elastin fibres intermixed with invasive tumour cells mixed with collagen fibres. that for an AIS case, with a size up to 16 cm, 150-300 formalin fixed paraffin embedded blocks had to be examined 14). Stromal changes in pulmonary adenocarcinomas Stromal changes in pulmonary ACs include accumulation of collagen fibres, increase in elastin fibres, or both, either as a mixture of both or as alternating layers. These go along with a marked reduction in airspace, which creates the sold appearance on high resolution of CT scan images. These localised stromal changes may or may not contain true invasive growth or alveolar filling growth. Palpation of the tumours with these localised stromal changes will feel firm, as opposed to the tumour areas without stromal changes, which are elastic. These changes have been called 'sclerotic' (3); a historical perspective on 'scar' carcinomas was published by Bobba et al (22). The initial notion of'scar carcinoma' Scanned with CamScanner 102 Histopathology of pulmonary carcinomas revisited: Defining invasion... emerged in 1939, referring to carcinomas originating in pre-existing pulmonary scars, recognised at gross examination. It was hypothesised that regenerating pulmonary epithelium at the periphery of the scar had an increased susceptibility to undergo malignant transformation (23,24). According to this hypothesis, the scar comes first and the tumour arises in it. More detailed knowledge of the stromal response in cancer modified this hypothesis: it is now assumed that carcinoma-associated fibrosis (desmoplasia) forms as a reaction to cancer cell proliferation (25-27). This hypothesis is consistent with the interpretation of the pathogenesis of the fibrosis associated with carcinomas in other organs, notably in the breast, stomach, pancreas, and colon (28-30] and is now the generally accepted concept. However, this does not exclude the possibility that some carcinomas may have arisen in a pre-existing scar (30). The terminology used to describe the morphology of stromal changes varies in the literature [31]. In some Japanese studies, the term fibrosis is used to describe a mixture of increased elastin and collagen or even a predominant increase in elastin (9,10,25,27,32].A white paper of the International Association for the Study of Lung Cancer (IASLC) uses the term sclerosis, describing an area at gross examination with a more firm consistency than normal pulmonary parenchyma, but without specifying fibre content or septal architecture [3]. Weydert and Cohen [30] stated that "Reproducible and universal definitions of'scar, 'fibrosis' and 'elastic framework' are needed if these morphologic features are to be used by practising surgical pathologists in the everyday sign-out of lung cancer specimens, and this still holds true today. In a study on small peripheral ACs with a scar Trejo Bittar et al [33] state that From a morphological standpoint, there is no histologic feature that could separate pre-existing subpleural scar/apical cap from scars associated with carcinomas. This suggests that what they recognised as scar was probably increased and condensed elastin without much collagen, the latter is in scar associated carcinomas in other organs the usual histologic characteristic [33]. We defined fibrosis as an increase in collagen fores, and this included a description of normal and increased elastin [7]. In a series of studies reporting on 'scar' [25,33-36) and 'morphology of AC' [9,11,32,37), a subgroup in the examined cohorts of ACs consistently showed tumours with preserved elastin framework, occasionally with increased elastin, but without clear invasion or desmoplasia, along with 100% 5-year survival. This subgroup likely corresponds to the cohort radiologically characterised by >50% ground glass opacities, morphologically classified in the WHOCOPA as AIS, minimal invasion and lepidic predominant AC. Some of these authors used the term "in-situ carcinoma' for this group. These literature data suggest that in the AIS category, lesions with collapsed AIS but also biological collapse without other signs of invasion should be included. Effect of tangential cutting Tissue sections of the lung cut through alveoli randomly. In a tissue section, a monolayer of cylindrical tumour cells on an alveolar wall will be oriented more or less perpendicular to the alveolar basement membrane, but due to the complex architecture of the lung there will also be numerous tangential cuts through alveoll lined with tumour cells In a tangentia cut, a monolayer might look like a multilayer; as shown in Figures 5.3 and 5.5b (14). AS a consequence, even though AlS is biologically a single layer of tumour cells growing on an alveolar wall, tangential sectioning will result in focal appearance of what looks like multilayering. This should be distinguished from real multilayering: the acquisition of the biological capacity of the tumour cells to grow in multiple layers. Pseudo-multilayering Scanned with CamScanner Multiple pulmonary tumour nodules 103 due to tangential cutting In AlS often shows three cell layers more than two cells wide, which satisfies the minimum criterion for micropapillary carcinoma according to the WHOCOPA, and hence overlaps with a criterion for invasion [38). Tangential cutting is also frequently seen in low-grade tubular adenoma of the colon, where foci within an adenomatous gland might appear pseudo-solid, which should not be considered as high grade (39). It is essential not to attach the weight of invasion to a multilayered appearance in tangentially cut AlS. A useful rule of thumb is that multilayering due to tangential cutting is focal within an alveolar space, and generally does not Involve the entire circumference of the alveolar wall. Tangential cutting may also result in seemingly free tumour cells in alveolar lumina in the plain of sectioning. This might suggest the presence of micropapillae, but should neither be interpreted as micropapillary growth nor as spread through air spaces. Extensive epithelial proliferation A recent study on tumours with a lepidic growth pattern attempted to identify histologic features, which may be used to establish criteria to distinguish between invasive and noninvasive areas 5). A Delphi process identified iatrogenic collapsed alveoli and transition zones with extensive epithelial proliferation as morphologic features prone to interobserver variability. The study ended with the suggestion to further examine these features. We then conducted a large reproducibility study which showed that iatrogenic collapse and biologic collapse are more important confusing features than extended epithelial proliferation (14). Extensive epithelial proliferation includes all intra-alveolar proliferations with two or more layers of tumour cells, including those falling short of the conventional micropapillary, acinar and solid patterns. For a zone with real multilayering of tumour cells, but not fulfilling the criteria for conventional micropapillary, cribriform and solid growth patterns, we have used the term 'grey zone' [14) or indeterminate. These morphological considerations form the basis of modified classification criteria which include: • The effects of iatrogenic collapse • The effects of biological collapse • Defined criteria for intra-alveolar growth (a minimum of three alveoli to avoid overdiagnosis due to tangential cutting) • The indeterminate category (grey zone) These modified criteria are summarised in Table 5.1. Features that are not informative for the decision of invasion are angulated or round glands without desmoplasia, thickening of the septa or inflammation (5). The characteristics of AlS and patterns of alveolar filling growth are related to various radiologic, gross, microscopic features and with importance for follow-up, as summarised in Table 5.2. MULTIPLE PULMONARY TUMOUR NODULES With the implementation of low-dose computed tomography screening, multiple pulmonary tumour nodules (MPTNs) are dagnosed with increasing frequency. This diagnosis should be considered when a benign process or metastatic disease from a cancer elsewhere are excluded. This diagnosis brings up the question of whether one is faced with Intrapulmonary metastatic (IPM) disease or with synchronous primary lung Scanned with CamScanner 104 Histopathology of pulmonary carcinomas revisited: Defining invasion... Table 5.1 Modified guidelines for the histological classification of adenocarcinoma 1A: AIS classification in situ (AIS), true invasion and alveolar filling growth Classification Histological hallmark Histological characteristics AIS (any) Regular monolayer of tumour cells, recognition of the alveolar wall supported by elastin stain • Regular CK7 staining pattern of tumour cells, borders of bronchovascular bundles and interlobular septa intact • Growth within alveoli in ≥3 adjacent alveolar spaces • Tangential cutting of a monolayer should not be interpreted as micropapillary or intermediate AIS with iatrogenic collapse Pseudopapillary and pseudoacinar appearance • Segmented elastin fibres in relatively thin alveolar walls without increase of elastin (Noguchi type A [40]) • Varying number of luminal alveolar macrophages, abundant presence signifies less iatrogenic collapse in that alveolus • In biopsies of the same lesion, the collapse may be less pronounced: undulating profile of alveolar walls • The effect of iatrogenic collapse may be less pronounced AIS with biological collapse when perfusion fixation has been used Increased elastin in • Linear elastin fibres sometimes very condensed (Noguchi alveolar wall type B) • Reduced alveolar size due to contraction • Minor collagen fibres may be present between elastin and tumour cells • Mixture with adjacent iatrogenic collapse possible 1B: Criteria for true invasion Modified criteria for true invasion Invaded structure Morphological characteristics Bronchovascular bundle Expansion in size, tumour cell growth underneath mucosai epithelium, in lymph or blood vessels; also visible in elastin stain Interlobular septum Tumour cells in septal area delineated by elastin on both sides Alveolar space (often deformed by desmoplastic stroma) • Tumour cells surrounded by desmoplastic stroma, focal or very extensive solid • CK7 stain supports recognition of individual tumour cells. Note that some pulmonary adenocarcinomas lack CK7 staining or have reduced staining intensity (41) Visceral pleura Tumour cell growth beyond pleural fibro-elastin layer, as in the WHO 2021[19) Alveolar wall stroma (interstitium) Frequently disrupted architecture with tumour cells surrounded by desmoplastic stroma and dispersed elastin. When alveolar walls are lined with alveolar type I! cells, metastatic disease should be considered CK7, cytokeratin 7; WHO, World Health Organisation Continues opposite Scanned with CamScanner Table 5.1 Continued 1C: Criteria for alveolar filling growth Modified criteria for alveolar filling growth Variants of alveolar Criterium filling growth Comments All variants Growth beyond a monolayer of tumour cells in ≥3 adjacent alveolar spaces This is a surrogate marker of invasion, often along with true invasion. Should be consistently present; in marked iatrogenic collapse, tangential cutting should not be confused with alveolar filling growth. With only alveolar filling growth (no lymphovascular invasion) the chance of recurrence is low Solid Complete alveolar filling Completely filled alveoli cannot collapse. In resection specimens iatrogenic collapse is seen in adjacent pulmonary parenchyma. alveolus in vivo The size of (filled) alveolus is that of a normal Cribriform Gland in gland formation (cribriform growth pattern) Overlaps with solid pattern. Alveoli filled with a cribriform growth pattern cannot collapse. In resection specimens, iatrogenic collapse is seen in adjacent pulmonary parenchyma Micropapillary Tufting and dissociated cells in alveoli. Shrinkage due to delayed fixation may detach tumour cells from underlying stroma, but often results in single cells. We do not favour the use of minimal invasive adenocarcinoma and the so-called invasive pattern of tumour spread through air spaces (STAS) May look like isolated detached tumour cells True papillary Fibrovascular cores without elastin in lumen of bronchi/bronchioli • Alveolar macrophages are usually absent • The original description of papillary carcinoma contains as an essential component. This may • Growth in alveolar space. Occasionally aiso desmoplastic stroma Indeterminate (grey zone) Multilayering more than in a tangential cut, but not enough to fall in one of the other subtypes represent true invasion Minor stratification of nuclei over the whole circumference of at least three alveoli. This has also been called'extensive epithelial proliferation (42). A maximally collapsed grey zone (<200 p) appears almost solid as opposite parts of alveolar walls are pressed together Multiple pulmonary tumour nodules 105 Scanned with CamScanner 106 Histopathology of pulmonary carcinomas revisited: Defining invasion... AIS' Lepidic? True papillary True acinar Micro- papillary Solid Type A3 Туре в* Lesion GGO"' Part solid" GGO/ Part solid Solid Solid GGO/Solid Solid Air amount >50% >50% >50% <20% <20% <50% Circumscribed/ Circumscribed <20% Edge Diffuse Diffuse Diffuse Circumscribed Circumscribed Diffuse Gross Palpation Soft/ Partly Flexible/ Firm Firm Flexible/Firm Firm Consistency Flexible flexible, Solid firm Easy to find No ⼟ Yes Yes ‡ Yes Yes Colour SAP SAP/ SAP White White SAP White/SAP White Edge Diffuse Diffuse Diffuse Sharp Absent Sharp Diffuse/Sharp Sharp Necrosis Absent Absent Possible Possible Possible Possible Histology Monolayer Yes Yes Yes/No Yes/No No No No Elastin in alveolar walls Thin short Increased I linear Short/ Increased No° No Yes Possible" Architecture regular Yes Yes Yes No No Yes/No No <10% Air amount? >50%- 2% >50%- 2% >50%- 2% <20% <20% <50% Transition to normal parenchyma Yes Yes Yes Possible Possible'° Yes'° Possible Alveolar filling Absent Absent No/ Yes! Yes No/Yes"3 Yes Yes growth"" True Absent Absent Focally Possible Possible Possible Possible invasion" Follow-up Recurrence No's No'S Possible16 Possible's Possible Possible's Possible Table 5.2 Characteristics of AIS and alveolar filling growth patterns are tabulated for radiologic, gross, microscopic features and follow-up chance Radiology SAP, similar as parenchyma 1 No size limitations for AlS. This implies that the whole lesion is embecided to exclude focal areas ofinvasion 2 Denotes growth on alveolar walls with at least focally true invasion. For this table, it includes indeterminate (grey zone). 3 AIS Type A is iatrogenic collapsed AlS with thin alveolar walls. With perfusion fixation the amount of'air in histology likely more. 4 AIS Type B is biological collapse. Mixtures of type A and B do occur. Multilayering in biological collapse is alved filing growth (surrogate marker of invasion) also elsewhere detectable. Continues oppos Scanned with CamScanner Multiple pulmonary tumour nodules 107 Table 5.2 Continued Similar for surgeons and pathologists. Not in tumour stroma, but present in pre-existing alveolar walls. Not surrounding of tumour formed glands, focal elastin fragments (destroyed alveolar wall) possible in tumour (desmoplastic) related stroma. Alveolar filling pattern may show underlying alveolar structure. Stroma when desmoplastic explains firm consistency. Depending on perfusion fixation or only submersion fixation. In the latter, less air. 10 Multilayering at edge. 11 Multilayering of tumour cells in three or more adjacent alveoli. 12 Invasive part may have alveolar filling. 13 Cribriform pattern. 14 True invasion in bronch(iol)us; pulmonary artery/vein; visceral pleura, lobular septum; desmoplastic stroma; true invasion with a size of 1 mm may metastasise. Category of' minimal invasive adenocarcinoma'is not used. 5-year disease-free survival 100% [5,43] 16 Patients with true invasion with or without alveolar filling growth have a recurrence chance of 25-30%. 17 According to the WHO classification (page 57 in citation [19]), AIS usually has ground glass opacity and sometimes also solid. However, tumour atelectasis may also give rise to a misleading solid appearance on high-resolution computed tomography (HRCT) [44]. Note that the radiology term'solid' has a different meaning than the pathology term'solid: 'white on HRCT versus complete alveolar filling growth' or'tumour cell invasion in desmoplastic stroma. The pathology correlates with radiologically solid (only implying that there is not enough air in the lung fitting with ground glass opacity) and ranges from inflammatory to benign and malignant diseases. cancers (SPLC). The distinction between these two is important for staging, management, and prognosis. Histological and molecular genetic analysis both contribute to making this distinction. A clonal relation between the nodules will denote metastases while the absence of a clonal relation provides a strong argument in favour of a second primary lung cancer. n lung cancer, three main histological types are discerned: (1) SCLC, (2) SqCC and (3) AC. Remarkably, the knowledge of (molecular) mechanisms that control histological types is limited. A frequent assumption in daily practice is that the morphology of a metastasis is usually similar to that of its primary 45). As a consequence, when two distinct intrapulmonary tumours are diagnosed, e.g. respectively as AC and SCC, this would constitute a strong argument in favour of two primary lung carcinomas, and this has been supported by molecular genetic analysis |46). However, in case of similar histological types, the likelihood of a clonal relationship is higher, but it is not unusual to find differences in acquired molecular genetic changes as we will discuss in the following paragraph (Figure 5.8). In 2009, Girard, Travis and colleagues introduced a comprehensive histologic assessment to differentiate between SPLC and IPM by comparing predominant and minor histologic subtypes between tumours, in addition to cytological and stromal features (47). In 2016, the staging committee proposed to add this to the ith Edition of the TNM classification for Lung Cancer (48,49). The question arises, however, whether extensive morphological analysis is really supportive in determining clonality. Although there is some (morpho) logic to this hypothesis, exceptions exist. Firstly, dedifferentiation will give rise to a change in morphology (50). Secondly, treatment may induce additional molecular abnormalities in tumour cells, leading to a marked morphological change, e.g. from AC to SCLC after epidermal growth factor receptor-tyrosine kinase inhibitor (EGIR-TKI) [51,52). In addition, Scanned with CamScanner 108 Histopathology of pulmonary carcinomas revisited: Defining invasion... RUL EGFR del exon 19 Three synchronous lung cancers RUL K-ras codon 13 RML K-ras codon 12 Figure 5.8 Patients with three synchronous lung tumours, where the morphology is indistinguishable in two tumours, while these tumours have a different driver mutation. Of note the morphological characteristics of the right upper lobe (RUL) and right middle lobe (RML) tumours do not support the use of comprehensive histologic assessment. a survey among pathologists revealed a high degree of uncertainty and disagreement in staging of cases with multiple lung carcinomas. Important in this context is the rarity ofa pure phenotype in pulmonary ACs {in a large series of 560 cases only three showed a single phenotype [53)}, which signifies that heterogeneity is inherent in ACs [54]. Furthermore, interobserver reproducibility in subtyping of ACs and poorly differentiated lung cancer has been found moderate to poor (4,55]. These findings point towards the need for improved histopathological criteria to distinguish separate primary tumours from intrapulmonary metastases. From a pathologist's perspective regarding consistency in making a diagnosis, a high posterior probability for a diagnosis, e.g. for distinction between SCLC and NSCLC, is important [56]. This holds true also for the differentiation between SPLC and IPM. We and others have performed large studies comparing histological with molecular classification [57,58). Some results are listed in Table 5.3. The probability of a 'correct diagnosis of IPM by comprehensive histological assessment, i.e. in line with molecular analysis, appears very low (around 60%). A level of 50% is like flipping a coin, and the 60% implies that in every three patients with MPTN diagnosed as IPM by histological subtyping on average in one patient, this is incorrect! We can conclude that the distinction between SPLC and IPM by histological criteria is a bridge too far, as illustrated in Figure 5.8. In contrast, the histological differentiation between AC and SCC, supported by immunohistochemistry on biopsies, is mostly consistent with molecular testing (57,58). Thus, restriction of the pathological analysis to histological typing remains reliable (64]. Molecular criteria The molecular tools used in the analysis of multiple lung tumours have included loss of heterozygosity (66-68), mitochondrial DNA |69), DNA microsatellite analysis, array comparative Scanned with CamScanner Multiple pulmonary tumour nodules 109 Table 5.3 Recent studies with comparison of predictive histology and molecular analysis for establishing the diagnosis of second primary lung cancer (SPLG) and Study Molecular intra-pulmonary metastases (IPM) First author classification Histological Case classification Comments [reference] number IPM SPLC Schneider (59] 60 IPM Asmar [60] Mansuet [61] Chang [62] Murphy [63] Vincenten [64] Yang (65] 69 109 76 41 107 SPLC IPM 7(48%) 8(52%) 12(46%) SPLC 14(54%) IPM SPLC IPM SPLC IPM SPLC 29(59%) 20(41%) 14(70%) 6 (30%) 11 (33%) 22(67%) Clonal 4(40%) 6(60%) 3(10%) KRAS, ALK, EGFR, MET 27 (90%) Maximum four genes (EGFR, KRAS, BRAF, ALK) n = 13 not informative; risk of over-interpreting same mutations as /PM 10(17%) 50(83%) 11(20%) 45 (80%) 0 (0%) 8 (100%) Mutation panel 22 genes Mutation panel 341-468 Non-clonal 72(97%) 2 (3%) 12(86%) 2 (6%) 31(94%) 2 (11%) genes Mutation panel DNA 11 genes; RNA fusion analysis four genes CNV; histological typing (no subtyping) 32 IPM Mutation panel 409 genes SPLC 2(14%) Indeterminate 0(0%) N, number of cases; max, maximal; CNV, copy number variation 15(83%) 1 (6%) genomic hybridisation [47,63,64,70,71), genomic breakpoint analysis (72), TP53, EGFR and KRAS mutation analysis (59,60,65,73-83), and next-generation sequencing (NGS) [84]. When a genomic difference is found at the level of a driver gene, this points towards a second primary tumour. In early studies on patients with MPTN's, a similar driver mutation was unequivocally interpreted as IPM (66,73-82,. In this context, acquired mutations in the TP53 gene show a broad spectrum with only a few, low frequency hotspots (85). Therefore, in contrast to KRAS and EGFR, mutation analysis of the TP53 gene is frequently effective in the examination of MPTNs (77,86). When similar mutations are found in KRAS and EGFR genes, the prevalence of the found mutation should be taken into consideration as these genes are known for hot spot mutations. The frequency of these differs between populations. In western countries KRAS mutations are found in approximately 26% of cases. For the hotspot mutations G12C, G12D and G12C (at 40%, 21% and 17%, respectively of all KRAS mutations) the prevalence is around 10%, 5% and 4.5%, respectively (87). In Asian countries the prevalence of KRAS mutation is around 12% 38). The distribution of the hot spot mutations is slightly different at for G12C, Gl2D and G12V, respectively at 30%, 18% and 15%. For EGFR mutations the estimated Asian and European prevalence is around 49% and 13%, respectively. In Asia and Europe, the frequency of exon 19 deletions is 49% and Scanned with CamScanner 110 Histopathology of pulmonary carcinomas revisited: Defining invasion... 48%, respectively, while for exon 21 1.858R substitutions, this is 41% and 30% (89). Thus, at population level, hotspot mutations in driver genes are not uncommon. For a patient with MPTNs, carcinogen exposure within the lung, and the genetic constitution of the xenobiotic metabolising enzymes will have been the same for all lesions. This implies that finding a hotspot (KRAS or EGPR) driver mutation in one tumour increases the chance of a similar hot spot mutation in another tumour. Therefore, the detection of a similar hot spot mutation in several lesions of a patient with MPTs is not definite proof for a clonal relationship. A better approach is to perform molecular tests with many data points, such as next generation sequencing covering many genes and shallow sequencing. When this approach detects several similar acquired genetic changes in both tumours, this confirms clonality and therefore IPM, while several differences (beside the common driver mutation) points to absence of clonality and therefore SPLC. How often this may be expected is shown in Table 5.4. Implications for staging Broad consensus exists that when, with all the approaches (histological typing, mutation analysis, LOH, CNV or extensive NGS) differences are found, a diagnosis of SPLC is justified. A simple schematic approach on how to handle MPTNs is shown in Figure 5.9. In practice, after histological typing we use mutation analyses on a limited number of genes Table 5.4 Individual cases with same driver mutation in two pulmonary tumours and different molecular finding with more extensive analysis First author [reference] Driver gene Additional gene/technique Number of Tumour 1 Tumour 2 Tumour 1 Tumour 2 cases/total cases Yang [65] Wild type 1/4 KRAS/TP53 KRAS KRAS TP53 codon 12 TGT codon 12 TGT codon 213 CGG Sozzi [90] KRAS/LOH KRAS KRAS Loss 3p24 No Loss 3p24 1/3 codon 12 codon 12 EGFR/KRAS EGFR L858R EGFR L858R 1/36 Takamochi [81] Arai [71] KRAS Wild type codon 12 AGT CGH concordance rate 2.6% De Bruin (91] Vincenten? 164] Chen [92) CGH EGFR EGFR Exon 19 del? Exon 19 del WES/WGS? EGFR L858R EGFR L858R No shared silent mutations CGH/ KRAS p.G12C KRAS p.G12C Non-clonal KRAS/EGFR 1/12 1/7 1/7 EGFR L858R EGFR L858R EGFR L858R EGFR L858R EGFR L858R' EGFR L858R' TP53 p. V272L Wild type TP53 p.E287' Wild type ERBB2 FGFR3 cndel' 3/20 Pei [84] Panel 808 EGFR L858R EGFR L858R en_amp' Higher unique/total mutation ratlo (>75%) 4/30 1 One case same EGFR (L858R) different FGFR3/ERBB2 (61] 2 WES, whole exome sequencing; GS, whole genome sequencing; CGH, comparative genomic hybridisation 3None of the metastases were studied at molecular level. Scanned with CamScanner Conclusion 111 Multiple pulmonary tumours Pathology. histological type Different Different Figure 5.9 Algorithm for analysis of multiple metachronous or synchronous pulmonary tumours to differentiate between independent primary tumour or metastasis. Different driver mutations are 100% reassuring, while similar driver mutations may with extensive molecular analysis still have a not neglectable chance of being synchronous primary lung cancers (SPLC). IPM, intrapulmonary metastasis Different * SPLC Same Mutation analysis Same Extensive molecular profiling Same IPM (or small NGS panel); when a difference in an acquired driver mutation is found SPLC can be concluded. When the same driver mutation is found, a larger panel is needed. The question then arises, when in the MPTN case two lesions share the same driver mutation but differ in other parameters, how many additional differences are required for a reliable call of independent primary. Wang et al and Pel et al suggest that if more than three common alterations are found, the call for IPM can reliably be made (45,84). The reasoning in terms of tumour biology is that pivotal driver alterations are present early on during tumourigenesis and trigger linear clonal expansion, that may be specific for the histologic type [93,94]. Careful analysis by Yatabe and colleagues revealed that spatial heterogeneous distribution of EGFR mutation is extremely rare (95), supporting the lineage of a driver gene as common ancestor. CONCLUSION We conclude that finding discordant driver mutations with various molecular approaches in a patient with MPTNs consistently point towards independent primary tumours (96,97). In contrast, when similar driver mutations are detected, this may signify metastatic lesions from a single primary. A 10% chance remains, especially when the decision hinges on high prevalence of hotspot mutations, that extensive molecular analysis examining more datapoints (large NGS panels, shallow sequencing, whole genome or exome sequencing). might yield convincing arguments in favour of a second primary tumour: Key points for clinical practice •Iatrogenic collapse leads to reduction of air, blood and lymph in the pulmonary resection specimens. • Deflation of alveoli leads to compensatory folding of the pre-existing alveolar wall, which may produce a papillary-like and an acinar-like pattern. • In collapsed AIS, these papillary-like and acinar-like patterns overlap with the WHO definition of invasive acinar and papillary carcinoma; this explains the poor reproducibility of histological assessment of invasion in pulmonary adenocarcinomas. Scanned with CamScanner 112 Histopathology of pulmonary carcinomas revisited: Defining invasion... • The modified classification incorporates (1) minimum criteria for alveolar filling growth and (ii) iatrogenic and biologic collapse. • Elastin and CK7 staining are helpful in recognition of collapsed AIS. Elastin in a fibrovascular core is a sign of a pre-existing alveolar wall. A regular pattern denotes growth on alveolar wall. • Alveolar filling growth is growth on the alveolar wall beyond a monolayer. • Histological typing may be helpful for distinction of morphologic differences in MPTs, but in case of similar morphology mutation analysis needs to be performed. • Different drivers or TP53 mutations are strong arguments in favour of different primary tumours. • As the same highly prevalent driver mutation may occur in different primary tumours, extensive molecular testing is advised when similar driver mutations are found. Frame ans of 20 marks for MD Pathology final year exams Make sure u do not miss any imp content