A Multidisciplinary Approach to Cancer: A Surgeon’s View

Introduction

Surgery remains a pillar of multimodality therapy for long-term survival in virtually all patients with solid tumors. For surgery to be effective, however, patients must be selected properly so that nontherapeutic surgery is avoided. In the current era of cancer surgery and imaging, “exploratory surgery” should, with the rarest of exceptions, not exist as a diagnostic modality. Both survival and resectability rates are rising based on many factors, including improved preoperative staging, but no test (biologic or radiologic) can tell the surgeon whether the patient should be operated. As an example, high-quality computed tomography (CT) imaging can predict degree of abutment versus encasement of the superior mesenteric artery with virtually 100% accuracy in pancreatic adenocarcinoma, yet other tumor–vessel relationships cannot be defined with 100% radioogical–­clinical correlative accuracy (e.g., hilar cholangiocarcinoma abutment or involvement of a sectoral hepatic artery), and small volume peritoneal disease may not be detected on even the best preoperative imaging, regardless of modality. These anatomic factors do not define whether a patient should be resected but rather what anatomical considerations define the resection planes. Clinically relevant underlying liver disease is inadequately assessed by imaging (steatosis, cirrhosis, hemochromatosis can be assessed and graded, for example), yet, despite radiological grading, the clinical relevance of underlying liver disease to a given planned procedure does not correlate well to imaging findings. Furthermore, the interaction between different treatments, including chemotherapy with and without biologically active agents, radiotherapy, intraarterial therapies, and surgery, require that treatment sequencing, timing, and duration be considered carefully. Together, these factors contribute to the constant movement in the line defining resectability for many tumors. Tumors (and liver parenchyma) may change character on imaging over time as some treatments are delivered, changing the sensitivity and specificity of radiologic findings as treatment progresses. Before patients embark on complex treatment plans, treatment sequence/timing issues must be considered by a team of physicians. Importantly, such multidisciplinary discussion and appropriate imaging must be completed before “palliative” treatments render potentially curable patients incurable (e.g., patients with resectable disease can be rendered unresectable if overtreatment with chemotherapy leads to liver toxicity). Open communication between surgeons and radiologists changes the way surgeons operate and changes the way radiologists report their findings to optimize patient care. Patient care is just that, patient care. If the focus of the radiologist, surgeon, radiotherapist, oncologist, and others in the care team is constantly on the patient, the best outcomes can be achieved. If the surgeon operates, the oncologist gives chemotherapy, and the radiation oncologist delivers radiotherapy based on a piece of paper (e.g., a radiology report), the best care may not be delivered. If the radiologist is integrated into the treatment team, modern, rapidly improving patient outcomes can be achieved more widely. Finally, goals of care differ in different patients. In some the goal is prevention, in others diagnosis and treatment. Imaging has a major role in screening and risk assessment in cohorts with a genetic tendency to develop hepatobiliary and other gastrointestinal cancers (e.g., BRCA1 and 2 for pancreatic cancer) or in assessing risk related to existing imaging findings (e.g., mucinous pancreatic cysts). In yet others, the goal may be palliation. Achieving these goals requires the members of the treatment team to work together in a patient-focused way.

Candidacy for “potentially curative” therapy is rapidly changing. As an example, multiple bilateral liver metastases from colorectal or neuroendocrine primary cancers can be treated with curative intent by surgery or by integration of interventional, percutaneous, and surgical approaches, leading to survival rates exceeding 50% at 5 years postresection. In such cases, radiology reports of “multiple bilateral liver metastases” may be accurate, but may also be misleading to the patient or oncologist/gastroenterologist reading the report (discussed later). Not every clinician is willing or able to review and understand imaging studies for every patient seen in the clinic. Surgeons should review every image on every patient, but primary care physicians, gastroenterologists, and medical oncologists are the gateways to surgeons, often encountering patients first and initiating treatment plans. Thus, clear assessment of imaging and clearly written reports are frequently the starting point for medical providers caring for patients with many tumor types. Multidisciplinary conferences help to overcome some of these problems because images are reviewed directly, but the vast majority of treated patients nationwide and worldwide will not be presented in multidisciplinary conferences. These factors contribute to the need for communication between members of the care team to optimize the value of imaging, as well as optimize patient care. This chapter will outline the following areas:

• Diagnosis

• Staging

• Surgical Planning

• Surgical Treatment

• Screening and follow-up in high-risk patients

Diagnosis

Accurate diagnosis may be based on patient history, clinical findings, imaging, biopsy or a combination of these elements. In rare cases, diagnosis may be made accurately with clinical history, tumor markers, and imaging, whereas pathology may be confusing (e.g., hepatic hemangioendothelioma, some cystic lesions in the liver and pancreas). Functional treatment modalities such as 2-[ ¹⁸ F] fluoro-2-deoxy-D-glucose positron emission tomography (FDG PET) and magnetic resonance imaging (MRI) can help clinicians in the correct clinical scenario but may also confuse the picture (by missing lesions in patients undergoing effective chemotherapy, and by highlighting nonmalignant areas of inflammation or postoperative change in other cases) unless the multidisciplinary team members work together to interpret studies in the proper context. Correct cancer diagnosis can often be made by review of imaging, depending on the disease site and type, as described in detail elsewhere in this book.

In many cases, however, pathologic diagnosis is needed, whether to confirm the clinical/radiological suspicion, as a requirement for treatment by radiotherapy or chemotherapy, or as a requirement for protocol-based therapy. In these cases, the initial imaging will often define whether a percutaneous or endoscopic approach to biopsy is needed. When percutaneous biopsy is planned, the presumed tumor type and location impact biopsy planning. For liver tumors, biopsy technique significantly impacts needle tract seeding, which should be an extremely rare event (<1%), even with primary liver tumors such as hepatocellular carcinoma. Ultrasound-guided lymph node biopsy with core biopsy can provide diagnosis of lymphoma, although excisional biopsy may be needed and is guided by both clinical examination and cross-sectional imaging; more recently, FDG PET may demonstrate the most active nodes and guide biopsy planning as well. Seeding is exceptionally rare with endoscopic ultrasound–guided biopsy and remains the standard of care for many tumors such as pancreatic and lower biliary lesions. Thus even at the level of obtaining diagnosis, consideration as to the probable diagnosis and possible treatments must often be given, reemphasizing the need for a multidisciplinary approach when treating cancer patients, including accurate prebiopsy imaging. Finally, interventions (percutaneous, endoscopic, or surgical) create artifacts that can impair quality of imaging; thus proper imaging before biopsy is critical for many reasons.

Staging

Once the diagnosis is made, the disease must be staged. Staging differs based not only on disease site but disease type, as treatments differ depending on findings (e.g., pancreatic adenocarcinoma with liver metastases is not treated surgically, whereas well-differentiated pancreatic neuroendocrine tumors with multiple bilateral liver metastases may be treated surgically with the expectation of very long-term survival); and the sensitivity of tests (and agents used such as FDG vs. gallium-68 for PET) differs based on disease subtype, such as mucinous versus nonmucinous gastric tumors, or tumor grade, such as well-differentiated versus poorly differentiated neuroendocrine tumors. Indications for chemo- and radiotherapy differ depending on the presence, absence, and often extent of distant disease (and sometimes imaging findings, as discussed later). Solid tumors are typically staged using the TNM system, although many elements of TNM staging are not typically assessed by cross-sectional imaging. T classification describes the primary tumor, and may relate to size (well-defined on cross-sectional imaging), but also may relate to depth of penetration through the layers of the organ (as with the gastrointestinal tract for esophagogastric, small bowel, and colorectal tumors), which is not evaluated on CT or MRI. Other factors such as invasion of adjacent organs and perforation may be seen well on imaging. Endoluminal MRI may classify invasion in rectal tumors, but otherwise CT and MRI are not used to assess whether layers of the gastrointestinal tract are invaded. N classification consistently relates to nodal involvement, although the N classifications differ from disease to disease based on the number and location of suspected or known nodal metastases. Imaging is central to further assessment of suspicious nodes based on nodal size, location, enhancement, diffusion restriction, and metabolic activity, and all contribute to surgical planning. M classification relates to metastases in all cases. Suspicious findings on any imaging study may be pursued, again given different impacts of metastases based on disease type. Treatments for different diseases with the same T, N, or M classification differ widely. Thus coordination between the imaging radiologist, interventional radiologist/gastroenterologist, and treating physician helps to ensure that the needed staging information is conveyed (and that the proper staging studies are obtained) to facilitate optimal patient care. In addition to classical staging information, surgical planning requires assessment of tumor–vessel relationships and anatomical variations, as discussed later.

Surgical Planning

Surgical planning depends on more than staging per se . Tumors in different locations are approached differently, and information about tumor–vessel and tumor–organ associations may not simply define resectability, but enable proper surgical planning:

• Tumor location/extent

• Tumor–vessel relationships

• Tumor–organ relationships

• Anatomical variations

Two different examples are described here. First, in the case of pancreatic adenocarcinoma, whether in the head, body, or tail of the pancreas, critical anatomical relationships define surgical resectability, including vascular abutment, vascular encasement, and occlusion by the tumor. Vascular involvement provokes significantly different treatment approaches (these are discussed in subsequent chapters and summarized here). Further, the vessel involved is important––arterial (superior mesenteric artery, celiac trunk, hepatic artery, splenic artery, left gastric artery, accessory, or replaced hepatic arteries) and mesenteric/portal venous involvement have significantly different surgical and oncologic implications. In the case of no vascular abutment or encasement based on multiphasic thin-cut CT or MRI, straightforward pancreatectomy is generally planned, with pre- or postoperative chemoradiotherapy. In the case of venous involvement, an entirely different operative plan is made to include vascular resection and reconstruction; this picture can change with neoadjuvant therapy. Venous abutment and encasement may not exclude resectability, whereas venous occlusion is considered “borderline” and may be resectable in selected cases. In the case of arterial involvement, preoperative therapy is generally advised, and in the case of extensive involvement (>180 degree encasement of the artery) surgery may not be indicated (again, this is a moving bar: the superior mesenteric artery and celiac artery are resected and replaced with venous and synthetic conduits with increasing frequency and improving results, so reports providing degree of involvement/encasement are most helpful to surgeons). The finding of liver metastasis virtually excludes the therapeutic value of surgery, even for the smallest resectable primary pancreatic adenocarcinoma. Regional adenopathy does not exclude resectability. Clear peritoneal carcinomatosis is a contraindication to surgery and to locoregional therapy; suggestion of peritoneal disease, which may not definitive (e.g., indistinct nodularity of peritoneum, mesenteric stranding, trace ascites) may prompt staging laparoscopy. Thus accurate staging and reporting of findings relevant to surgery for the specific disease are critical to surgical planning and result from communication between imaging and treating physicians. Primary tumor type impacts the requirements of the report––primary pancreatic adenocarcinoma with a single liver metastasis is not a surgical disease; primary colorectal cancer with 17 bilateral liver metastases may be treated with curative intent, as described later.

As a different example, issues in liver surgery can be even more complex. Resectable liver tumor(s) are often defined based on liver that will remain after resection, including preservation of adequate inflow and outflow to the preserved segments, with adequate liver remnant volumes. Tumor–vessel relationships within the liver impact resectability differently than for pancreatic or other gastrointestinal, thoracic, head and neck, or extremity tumors. Vascular resection (portal, hepatic venous, caval, and arterial) is increasingly performed in many centers, depending on disease type and often therapy response. Tumors may involve two of three outflow vessels (hepatic veins) in the liver and abut the inferior vena cava but be resectable with standard techniques and excellent results. Rarely, involvement of all three hepatic veins, traditionally considered a sign of unresectablity, is not an impediment to complete resection either, because vascular resection/reconstruction can be considered, or because venous anomalies may permit otherwise impossible resections such as subtotal hepatectomy based on the presence of a dominant inferior right hepatic vein. Major hepatectomy includes resection of tumors involving major hepatic and portal branches routinely, as long as vessels supplying and draining the liver remnant are free of tumor. In other cases, major resection is possible because of other anatomic variations in the liver, such as a staged portal bifurcation allowing resection of tumors involving the central liver. Arterial and biliary anomalies contribute to assessment of resectability in a significant proportion of patients.

Thus surgeons and radiologists must understand the anatomy of the liver (the hepatic veins, portal veins, and hepatic arteries) and must remark variations including replaced or accessory hepatic arteries (present in up to 55% of patients) or even the existence of important venous variants (e.g. inferior hepatic veins). Segmental liver volume is highly variable, which impacts surgical planning ; and tumor involvement of intrahepatic vessels or portal structures, as well as therapy changes, can alter volume distribution among hepatic segments. Systematic liver volumetry based on cross-sectional imaging and/or functional imaging is a critical tool to surgical planning for major liver resection, reiterating the intersection of radiologists and surgeons in surgical planning, although liver volumetry is typically requested by the surgeon to facilitate surgical planning that leaves specific adjacent anatomical liver segments intact. Radiologists and surgeons who work together are aware of the importance of anatomic variations and tumor–vessel relationships, and that these factors can lead to different radiological reports that guide patients and clinicians with the help of high-quality imaging and interpretation. An example synthesizing these issues in a patient with multiple bilateral colorectal liver metastases is illustrated in Fig. 2.1 .

Finally, advances in imaging have advanced the correlation between imaging findings and patient outcomes. Two examples warrant comment, both of which relate to treatment of solid tumors with “biologic” agents and assessment of response on CT (and/or PET). The first is gastrointestinal stromal tumor (GIST), which treated with imatinib mesylate, an inhibitor of KIT and PDGFRa tyrosine kinases. Traditional methods of assessing response such as Response Evaluation Criteria in Solid Tumors (RECIST) are not always sufficient to capture the effects of this class of new agents, which cause less size change and more cystic change and loss of vascularity in GIST tumors than traditional treatments, leading to a shift in assessment of response in these tumors that is based on different radiologic criteria. Even in colorectal liver metastases, RECIST is a poor indicator of response to newer agents such as bevacizumab, a vascular endothelial growth factor antagonist. One study has shown that morphologic criteria that focus not on tumor size changes but on changes in vascularity and the margin between the tumor and the liver predict survival in patients with resectable and unresectable colorectal liver metastases treated with bevacizumab, and that the morphologic radiologic response (but not RECIST) correlates with pathologic response to chemotherapy, a true survival predictor. MRI with diffusion-weighted imaging may also provide information as to the degree of activity of liver metastases after chemotherapy (or intraarterial therapy). MRI may have an advantage in lesion detection after chemotherapy as well, owing to changes in the character of the tumor and the liver impacting the “contrast” between lesion and normal underlying liver. These examples of advances in imaging, as well as the correlation between newer imaging findings and outcomes, are important to physicians and surgeons who determine treatment plans. At the same time, shortcomings of even the most modern imaging techniques must be considered. The absence of PET activity or arterial enhancement of a GIST or colorectal liver metastasis almost never (<10%) indicates cure, so again, oncologists, surgeons, and radiologists must avoid overinterpretation of findings on imaging before treatment decisions are made.