The more we, the medical community, learn about the biology and treatment of metastatic cancer, the less effective we appear to be at predicting the survival of patients afflicted with the disease. Prior to President Nixon’s “War on Cancer” (launched in 1971), survival of patients with MBD was usually predictable. With few effective chemotherapy agents and no targeted therapies (with the exception of hormonal modulation/ablation), survival was predictably short for most patients [12]. Improved treatment has resulted in the fortuitously increased challenge of life expectancy prognostication for many types of metastatic cancer. A central component of informed consent is discussion of the risks, benefits, and alternatives of the proposed intervention [13]. Benefits of intervention depend upon at least two variables: the improvement gained from the intervention and the duration for which the improvement lasts. Patients with limited life expectancy may not benefit from well-intentioned MBD surgery if the requisite recovery period is longer than their remaining survival. It would appear logical that extensive clinical experience or the pooled expertise of multiple providers should improve prognostication. Unfortunately, a 2007 study found that an experienced clinician was no better than trainees (average 1.7 years experience) at prognostication (<1 month, 1–6 months, >6 months) in a prospectively evaluated cohort of 216 metastatic cancer patients. Furthermore, a multidisciplinary tumor board was not better than the trainees or the individual experienced physician [14]. Survival was systematically overestimated among the 15 % of patients who died within one month. Another study analyzed 395 predictions made by 8 different disciplines within a radiation oncology unit. Overall, survival was overestimated 72.4 % of the time. Radiation oncologists, radiation therapist, and nurses had equivalent predictive accuracy in a study comparing physician, nursing, allied health professional predictions of survival [15]. Accuracy of prognostication varies with the time frame being predicted. A study of primarily general surgical oncology patients undergoing palliative procedures found that surgeons could accurately predict survival of less than or greater than one year. Surgeons also tended to underestimate the palliative benefit of surgery relative to validated patient self-reported questionnaires [16].

Due to the poor prognostic ability of physicians, investigators have sought to use patient derived data to make predictions. A study of 1636 patients with metastatic lung or colorectal cancer found that patients frequently self-report worse nutritional intake and a lower KPS than their oncologists. Patients had objectively worse survival when their self-reported KPS differed from that of their physician [17]. Another study of patient derived performance status showed good predictive ability of the patient derived data in three distinct cohorts: home hospice, inpatient, and ambulatory care [18]. While patient’s self-assessment has shown utility for prognostication, caution is warranted in evaluating patients with high levels of psychosocial distress. A study of over 1000 terminal cancer patients failed to identify psychosocial distress as an accurate predictor of survival although its identification may impact the patients’ treatment plan and overall well-being [19]. The Palliative Prognostic (PaP) score was created in an effort to objectively prognosticate survival of metastatic cancer patients. The PaP combines a clinician estimate of survival, KPS, presence or absence of dyspnea and anorexia, total white blood cell count, and lymphocyte percentage; it has been validated in several different cohorts [20]. A subsequent study found that inclusion of delirium improved the prognostic capability of the PaP [21].

Every patient does not seek an estimate of survival, nor do they always use the information in selecting a treatment plan. A 2005 Australian study found that only 61 % of patients wished to know their prognosis, and only 52 % of this subset wanted to know their exact (within weeks) prognosis [22]. Cultural differences exist with respect to the desire for prognostic information, and culturally competent approach can improve patient and family satisfaction [23]. The data summarized above suggests that patient derived data, and the use of a validated scoring system can improve prognostic accuracy. Regardless of the prognosis or its accuracy, survey data indicate that patients appreciate physicians who appear honest and competent and dislike the use of euphemisms and physicians who appear uncomfortable or nervous in discussing end of life issues [24].

The urgency of surgery is difficult to define as reflected by the updated ACA/AHA guidelines discussed above. Delay of 1 week versus 6 weeks (the range defining time sensitive cases by ACA/AHA) may be inconsequential for many solid organ tumors requiring surgery, but few orthopedic surgeons consider this time differential equivalent when addressing a fracture. Threats to life and limb, severe pain, and resource availability and allocation may all impact the timing of surgery. Most pathologic fractures are low energy, mildly displaced, and associated with less soft tissue injury than the fractures due to high energy polytrauma encountered commonly by the majority of orthopedic surgeons. Despite this fact, pathologic fractures may rarely cause hemorrhage, severe soft tissue damage, and systemic inflammatory response necessitating more of a “trauma” as opposed to an “oncologic” strategy. In reality, this bimodal view is myopic as patients present on a continuum necessitating the use of best practices from traumatologic and oncologic orthopedics.

A common timing dilemma occurs when an impending or non-displaced pathologic fracture occurs in a patient receiving chemotherapy. In such cases, acute surgical intervention may compromise their systemic treatment or place them at increased risk of postoperative complications such as infection. In these cases, the consequences of displacement of the fracture upon the surgical plan should be considered. Femoral head and neck lesions are typically treated with arthroplasty; furthermore, the complexity of the arthroplasty is generally not increased by displacement of a femoral neck fracture. In contrast, non-displaced fractures of the intertrochanteric and subtrochanteric femur may be treated with a simple intramedullary nail whereas their displaced counterparts may require complex primary arthroplasty or even proximal femoral replacement [25]. As such, the consequences of waiting are far more severe for the latter than the former despite the difference in location within the bone of mere centimeters. Upper extremity lesions may also be treated more expectantly as even displaced fractures are unlikely to result in a bed bound state or necessitate inpatient hospitalization [26]. Most upper extremity pathologic fractures may be temporized with sling or splint immobilization while the patient is optimized for surgery. Medical oncologists must rely upon their orthopedic colleagues for discernment of these subtle nuances. A true multidisciplinary team effort is frequently required to obtain the most satisfactory solution.

Most patients receiving cytotoxic chemotherapy experience neutropenia which is commonly defined as an absolute neutrophil count (ANC) of less than 1500 neutrophils per microliter with severe neutropenia defined as an ANC less than 500 [27]. Metastatic cancer is an independent risk factor for surgical site infection, and neutropenia further increases this risk [28]. Severe neutropenia increases the infection risk of even small implant procedures such as vascular access port placement [29]. The timing of the ANC nadir will vary with the specific agents being used and the patient’s general condition and past response to chemotherapy. Close coordination with the medical oncologists is required to avoid scheduling surgical intervention during a period of neutropenia. There are no absolute guidelines for delaying surgery in the neutropenic patient. One study of intra-abdominal surgeries in neutropenic patients concluded that waiting for neutrophil recovery if at all possible was most prudent [30]. It is the author’s preference to delay surgery until the ANC is >1000 unless the clinical scenario absolutely dictates otherwise.

Targeted therapies have revolutionized the treatment of many malignancies including those frequently metastasizing to bone (Table 19.2). While neutropenia is less frequent with these agents than cytotoxic therapy, the effects of targeted agents on wound and bone healing are largely unknown. Bevacizumab , a vascular endothelial growth factor (VEGF) inhibitor, has well-documented adverse effects on surgical wound healing [31]. VEGF is a key regulator of angiogenesis which is crucial for both tumor growth and wound healing. In addition to generalized wound complications, it has specifically been shown to impair bone healing after craniotomies for brain metastases [32]. The majority (60–80 %) of patients treated with epidermal growth factor receptor (EGFR) inhibitors develop skin toxicities, and 38 % of patients with skin pathology developed superficial infections in a study of 221 patients [33]. The majority of these infections were Staph Aureus with 5 % being methicillin resistant. Human epidermal growth factor receptor 2 (HER2) is EGFR family member particularly important in breast cancer. In addition to skin toxicities, the HER2 antagonists cause left ventricular dysfunction in up to 18 % of patients [34] which may be of particular relevance for planned surgical interventions. Knowledge of the risks of all selective therapies is unrealistic; however, orthopedic surgeons should be aware that significant perioperative risks which may not be identified with routine preoperative testing exist with these agents. Close communication with the treating medical oncologist is important to both fully inform patients of the attendant risks and to potentially mitigate the risks.